Applied Cancer Screening Research Branch, Behavioral Research Program, Division of Cancer Control and Population Sciences, National Cancer Institute, National Institutes of Health, Rockville, Maryland
Applied Cancer Screening Research Branch, Behavioral Research Program, Division of Cancer Control and Population Sciences, National Cancer Institute, 6130 Executive Boulevard, Suite 4102, Rockville, MD 20852
This article is a US Government work and, as such, is in the public domain in the United States of America.
This article provides an overview of behavioral and social science cancer screening intervention research and introduces the scope of topics addressed in this supplement to Cancer. The authors identify and address issues to consider before conducting interventions to promote the uptake of screening tests, such as the benefits and harms associated with screening. Trends in the use of cancer screening tests are discussed in the context of their efficacy and adoption over time. Both the development and breadth of social and behavioral intervention research intended to increase the use of effective tests are reviewed as background for the articles that follow. The application of the lessons from this extensive knowledge base not only should accelerate the uptake of the effective cancer screening tests currently available, but also can guide future directions for research. Cancer 2004. Published 2004 by the American Cancer Society.
Evaluating the effectiveness of different strategies to promote cancer screening has been the focus of research for several decades. The accumulated literature (which now includes numerous meta-analyses and literature reviews) chronicles many effective approaches to increasing the uptake of screening1–9 as well as strategies that intuitively seemed promising but in the end were not.10 Our challenge now, and the purpose of this supplement to Cancer, is to identify lessons learned from experience in conducting intervention research and related studies aimed at improving the use of tests recommended for the early detection or even prevention of cancer. In doing so, we hope not only to encourage the application of evidence-based interventions but also to refine our approaches to promote cancer screening by stimulating researchers to consider how the lessons from this extensive knowledge base may translate into new cancer screening tests.
Intervention science focused on screening appears quite different now than it did when strategies first were introduced to promote screening for cervical cancer. Early efforts to promote cervical cancer screening, and later breast cancer screening, attempted to increase public awareness regarding the benefits of early detection and used community outreach strategies, such as mass media campaigns and promotional mailings.11–18 As the use of screening with mammography became more common, and emphasis increased concerning screening for other cancers, the environment in which health services were delivered changed, as did communication technology and health care consumers themselves. Research over time identified the pivotal role of physicians and other health care professionals in facilitating screening.19, 20 In addition, structural influences (including policy and socioeconomic factors, health insurance coverage, and access to a regular source of health care) were recognized as either facilitating or impeding screening.21 In the early 1980s, it was assumed that if screening for breast cancer were promoted, and especially if free or low-cost mammograms were offered, most women would be screened. However, events proved otherwise, and the low uptake of mammography became a catalyst for more than a generation of behavioral research.
Initial research focused on identifying women's barriers to screening22; soon, researchers began to develop and test interventions to overcome these barriers. Over time, investigators recognized that the barriers were multidimensional, and therefore the interventions should be as well. The growing awareness of the influence of individual as well as structural factors led investigators to test multicomponent interventions that simultaneously addressed multiple levels (e.g., patients, providers, organizations, and communities). Today, it should be possible to accelerate the uptake of effective screening tests by applying what already has been learned in a more strategic and integrated manner.
The first article in this supplement introduces the scope of intervention research and related studies concerned with improving adherence to cancer screening tests. We first provide background concerning the basic principles of screening and identify issues that are important to consider before embarking on interventions to promote the uptake of tests. Next, we provide an overview of current trends in cancer screening and of intervention research intended to increase the use of effective tests. We conclude with an outline of the topics addressed in the supplement.
THE FUNDAMENTALS OF SCREENING
This supplement to Cancer focuses on social and behavioral research, and specifically intervention-related research, to enhance adherence to screening for cancer. Although the outcome under study is “screening behavior,” it is important to understand what constitutes sufficient evidence to justify promoting a screening test. Because of the financial, social, and psychologic costs associated with screening, both to individual people and society, the benefits of screening should outweigh the costs before the use of a given test is promoted. Evidence regarding the efficacy of a test is not the only factor influencing screening behavior, but it is nevertheless inextricably linked to prevalence in use. Because evidence to support the use of screening tests often is ambiguous, messages communicating the advantages of screening also are complex. This complexity poses unique challenges for designing interventions to influence cancer screening behavior.
Designing, conducting, and interpreting evidence from studies evaluating the efficacy of screening tests presents its own set of challenges23; a detailed discussion of this topic is beyond the scope of this article. Herein, we review the basic principles of screening, the complex issues that must be considered before deciding whether to promote a given screening test, and reasons why policy decisions regarding screening may be controversial.
Potential Harms Associated with Screening
The potential harms of screening range from mild to severe, and their relative importance may differ based on the perspective of those evaluating the test (e.g., patients, providers, or health plans). The absolute risk of harm may be small, with only a very small proportion of screenees likely to be affected, but some screening tests, such as colonoscopy, can have serious consequences (e.g., perforation of the colon). Furthermore, because the specificity of all currently available screening tests is < 100%, screening will produce at least some false-positive results. Conversely, because sensitivity is always below 100%, the harm of false-negative findings must be considered, such as having the patient ignore symptoms or, if a tumor is detected at a later date, losing trust in screening. Patients with false-positive results may experience unnecessary anxiety and possibly endure costly and invasive medical intervention. Another potential harm to consider is the diagnosis of clinically insignificant conditions as screening tests become more sensitive in detecting very small tumors or those with less potential to be invasive. Because our ethical responsibility is to avoid harm when screening healthy, asymptomatic populations, we must promote only those tests in which the scientific evidence demonstrates that the benefits outweigh the harms and are sufficient to justify the human and financial costs. This has been the approach of the U.S. Preventive Services Task Force (USPSTF); we support the principles under which the Task Force operates (available from URL: http://www.ahcpr.gov/clinic/prevenix.htm [accessed 7/6/2004]).
Principles of Screening
The decision to screen an at-risk population for cancer is based on well established criteria.23 Although the overall goal is to reduce morbidity and mortality from cancer, the objective of screening per se is the use of a relatively simple, inexpensive test in a large number of individuals to determine whether they are likely or unlikely to have the cancer for which they are being screened.23 This principle is commonly cited, but many of the currently available screening tests are neither simple nor inexpensive. The emphasis on classifying people on the basis of their likelihood of having cancer is very much the point of cancer screening, because screening tests are not diagnostic tests. Rather, they should distinguish those individuals who may have the disease from the much larger group of people who most likely do not have it.
Before it is recommended that large numbers of healthy persons with no signs or symptoms of cancer undergo routine screening, certain criteria, first promoted by the World Health Organization, are commonly used to assure a net benefit to screenees.24, 25
The disease in question should be an important health problem; its significance may be defined by several criteria related to disease burden, including morbidity, mortality, and premature mortality.
There should be a detectable preclinical phase (i.e., a period in the natural history of the disease when it is detectable in asymptomatic individuals, thus gaining lead time before symptoms develop).
Treatment of occult disease (i.e., disease diagnosed before the symptoms appear) should offer advantages compared with the treatment of symptomatic disease.
A screening test should be affordable and provide benefits justifying its cost.
The test must be acceptable to the target population and to health care professionals.
Screening tests must achieve an acceptable level of accuracy in the population undergoing screening. The earlier emphasis on the likelihood of disease also is important in this context, because screening tests are never 100% accurate. Accuracy varies by test, cancer site, and individual patient characteristics. To our knowledge, no threshold criteria exist for an acceptable rate of false-positive and false-negative results.26 Instead, issues related to test accuracy should be evaluated in terms of the relative consequences of one or the other kind of error.
The ultimate goal of a cancer screening program is to reduce site-specific mortality. Screening helps reduce morbidity and mortality by providing an opportunity to initiate treatment earlier in the natural history of a disease. Some screening tests (e.g., the Papanicolaou [Pap] test, colonoscopy) also may prevent cancer by identifying precursor lesions (i.e., dysplasia, polyps) that, if removed or treated, reduce the likelihood of developing invasive disease. In the absence of effective treatment, or in the case of a rapidly progressing disease, screening is of no value.
At the most fundamental level, the goal of screening is to improve outcomes. If treatment before the onset of symptoms is beneficial, then improvements in survival should be associated with the gain in lead time, and mortality should be lower in those patients diagnosed by screening compared with those diagnosed after symptoms are apparent. Conversely, if the gain in lead time advances the time of diagnosis, but the patient's life is not extended beyond when death would occur without early detection, there is only the appearance of greater survival. Moreover, those patients who are diagnosed are living with an awareness of cancer—a fact that may itself carry negative sequelae.
To our knowledge, randomized controlled trials (RCTs) provide the strongest evidence to support the efficacy of a given test because they reduce many known biases through random assignment. However, it is not always possible to conduct RCTs and they have limitations. For example, RCTs may not be effective in estimating the true magnitude of the screening benefit because of nonadherence in the group invited to screening as well as contamination (i.e., participation in screening) among the “not invited,” or usual-care group, thus underestimating the true magnitude of benefit among screened individuals.
Multiple organizations have established systematic review processes for evaluating evidence of efficacy from cancer screening trials, including the USPSTF (available from URL: http://www.ahrq.gov/clinic/prevenix.htm [accessed 7/6/2004]), the Cochrane Review (available from URL: http://www.cochrane.org [accessed 7/6/2004]), the American Cancer Society (available from URL: http://www.cancer.org [accessed 7/6/2004]), the National Institutes of Health through its Consensus Development Conference reviews (available from URL: http://consensus.nih.gov [accessed 7/6/2004]), and others. The reviews provide a foundation for governmental, voluntary, and professional organizations to issue recommendations and guidelines for the use of cancer screening tests. Because methodologic approaches to conducting evidence review may vary, it is not uncommon for organizations to interpret trial data and other evidence differently. Nevertheless, most would agree that a cancer screening test should not be promoted before establishing its benefits.
Implications of Evolving Science
Keeping pace with the changing landscape of technology and evidence presents challenges for communicating the state-of-the-science to health professionals, policy makers, and the public at large. In the absence of definitive RCT evidence, new technologies may be promoted as overcoming the shortcomings of previous technology, but new is not always better. Even if good evidence exists of a benefit from screening based on old technology, new technology still requires careful evaluation. For example, new technology may offer improved sensitivity but poorer specificity, or specificity may be improved but sensitivity may be the same. The new test may be equally accurate, but cheaper; alternatively, it may offer modest improvements but at greater expense. Furthermore, conducting new trials that randomize one group to an invitation to screening and another to usual care is not, in general, ethically feasible after the benefits of screening have been established. For instance, in the 1990s, it was considered infeasible to conduct a new trial of mammography in women ages 40–49 years27 because a large proportion of the population already had adopted the practice.
These circumstances not only present challenges to evaluation and decision-making but also must be viewed against the backdrop of various interest groups, many of which are very vocal. Interpretation of new evidence or technologic advances will necessarily vary by perspective and vested interest. Individual patients will likely weigh the benefits and risks of undergoing screening based on their perceptions of risk, values, and other personal factors, whereas policy makers will look for enough evidence to justify the allocation of resources to reimburse for a procedure.
Debates regarding cancer screening tests often simmer among scientists and policy makers and occasionally boil over in the mass media. These debates highlight the challenges of developing interventions to promote utilization while simultaneously acknowledging the limitations of both the science and the test. Experience with mammography is illustrative. In this case, controversies began in the late 1970s but were most apparent in 1993 and 1997, when national panels reviewed the available data regarding when women should begin undergoing mammograms and the appropriate screening interval.28, 29 The predominant issue was whether women in their 40s should be advised to undergo regular mammography. At the core of this debate were differences of opinion regarding whether the existing RCT data should be the sole basis for decisions concerning screening in this group, or whether inferential data as well as meta-analyses indicating a benefit should take precedence.30–33 A later controversy was ignited in 2001 with the publication of a Nordic Cochrane Center meta-analysis34 suggesting that evidence was insufficient to recommend mammography screening for women of any age because of alleged methodologic flaws in the trials. Several months later, that point of view was countered with reports from numerous expert groups, including the USPSTF, concluding that the Cochrane analysis had not identified quality limitations that were sufficiently concerning to overcome long-standing judgment regarding the value of mammography.35–39 To our knowledge, little evidence exists that conflicting scientific opinions affected overall consumer or provider intentions to continue screening,40 but this experience highlighted the fact that communicating a balanced perspective on screening may be just as important as, if not more important than, promoting the use of the screening test itself.
TRENDS IN THE USE OF CANCER SCREENING TESTS
Intervention studies and the widespread promotion of screening usually are conducted after achieving scientific consensus that screening improves health outcomes. Nevertheless, evidence is not the only factor driving the adoption of screening tests. Examining trends in the use of screening provides some insight into the many factors that influence adherence to the use of tests that have been widely accepted as efficacious. Figure 1 displays the prevalence and progress of cancer screening since 1987, when we began national monitoring of the prevalence of tests. What is striking is the different rate at which the tests have been adopted over time, particularly in view of the different levels of evidence for each test. From 1987–2000, the prevalence of cervical cancer screening in the U.S. remained consistently higher than screening for any other cancer. In part, this is due to the very convincing indirect evidence from observational studies that screening for cervical cancer reduces its incidence and mortality.41 The test also is relatively inexpensive, easy to perform, and accepted by physicians without controversy surrounding its benefits. As a result, cervical cancer screening has been widely adopted in many countries. The USPSTF recently recommended reducing the frequency of Pap testing because more frequent screening does not appear to identify substantially more cancers than less frequent screening,42 but the extent to which women and their providers will follow the new guidelines is difficult to predict. Although the prevalence of screening for cervical cancer is quite high in most developed countries, this is not the case in the developing world, in which cervical cancer is a major source of mortality.
In contrast to patterns in screening for cervical cancer, the late 1980s and early 1990s saw a dramatic increase in the use of mammography. Available clinical trial data in the 1980s demonstrating that screening mammography reduced mortality among women ages 50–69 years led many organizations to recommend regular screening mammography for women in that age group. Consequently, many interventions to promote the use of this screening test were implemented during the 1980s and 1990s. As discussed previously, the increasing trend in mammography use during the 1990s appears to have been largely unaffected by the controversies surrounding the test during that period.
The uptake of colorectal cancer screening, however, has been relatively low for both men and women (< 30% reported the recent use of any test in 2000). The first guidelines recommending screening for colorectal cancer, published in 1980, were based on persuasive inferential evidence of the effectiveness of fecal occult blood testing (FOBT) and rigid sigmoidoscopy.43 However, it was not until 1997 that evidence-based guidelines for screening could be drawn on the basis of data from RCTs and case–control studies.44, 45 Although there has been strong expert consensus regarding screening for colorectal cancer since the late 1990s, we have not yet witnessed the dramatic increase in use that was observed with mammography after the RCT data were published.46 However, the multiple options for colorectal cancer screening may both facilitate and hinder use because greater flexibility in adapting to patient preferences and the variety of test options may create confusion among patients and providers. In addition, a misperception that only one of the colorectal cancer screening exams (colonoscopy) is efficacious may prevent the use of other effective modalities (i.e., FOBT and flexible sigmoidoscopy) that are likely to be accessible to most people.47 In fact, to our knowledge, the only randomized trial data supporting screening are for FOBT.48 Finally, feelings of embarrassment surrounding colorectal cancer make communicating about this disease particularly challenging.
In contrast to the situation with Pap tests, mammography, and testing for colorectal cancer, there is a lack of data concerning the efficacy of screening for prostate cancer, but this has not prevented clinicians from recommending and performing the tests or men from requesting prostate-specific antigen (PSA) tests. In 2000, self-reported data indicated that 41% of men age ≥ 50 years had undergone a recent PSA test,46 but because physicians occasionally order PSA tests without obtaining active consent,49 the true proportion could be even higher. Decisions are particularly difficult regarding the trade-offs of detecting this cancer early. Some prostate cancers are slow growing and may not seriously affect health for many years or at all during a man's lifetime, yet others are virulent and lethal. Early intervention may prevent death from prostate cancer, but also might result in impotence or urinary incontinence, which impair quality of life. Because so many prostate cancers are not life-threatening, the potentially adverse consequences of treatment make informed decision-making regarding screening especially challenging.
The extent of expert consensus regarding the benefits of early detection varies for the different screening technologies currently available, depends in part on the standards of evidence applied, and is only one factor influencing how readily a given technology is adopted into clinical practice. Until recently, intervention research has focused largely on the promotion of screening tests with demonstrated efficacy, but the substantial variability in interpreting what constitutes “strong enough” evidence to endorse recommendations for most cancer screening tests has pushed research in new directions. Additional factors (e.g., cost, invasiveness, time required, benefits, harms) that are specific to each screening modality must be considered when any new test is introduced (or better yet, before it is introduced) because they will significantly shape approaches to intervention. Also, for some widely used screening procedures, there is little or no evidence of effectiveness (available from URL: http://www.ahcpr.gov/clinic/prevenix.htm [accessed 7/6/2004]). Discouraging the use of ineffective screening tests or promoting informed decision-making50 may be as challenging as attempting to increase the use of effective tests.
Contrasting trends in the use of cancer screening tests raise questions regarding the best ways to promote cancer screening in the future. For example, what elements of successful screening promotions consistently have proven relevant? Will the same strategies that successfully increased mammography rates ultimately produce greater use of colorectal cancer screening? Will these strategies apply regardless of changing technology or the introduction of new screening tests? How can we improve on prior intervention approaches, and where is further research needed? The articles that follow in this supplement address these issues by discussing the development, implementation, evaluation, and dissemination of strategies intended to influence cancer screening behavior and follow-up.
INTERVENTION RESEARCH IN CANCER SCREENING
Intervention research is informed by epidemiologic and clinical research linking screening to health outcomes such as morbidity and mortality, but intervention research is not intended to answer questions regarding the efficacy of a given screening test or technology. Rather, it addresses a dual need: 1) promoting more widespread use of effective screening while 2) helping consumers and providers understand the value and limitations of screening. Intervention research in cancer screening has long been interdisciplinary. The scope of this research involves more than testing the effects of a given strategy to change patient, provider, or organizational screening practices, because it also includes 1) studies to identify barriers to and facilitators of screening, 2) methodologic investigations, and 3) surveillance to distinguish the characteristics of a population or setting that can influence the design and implementation of a screening program. Intervention research also focuses on questions concerning how individual people respond to abnormal test results and ways to increase adherence to follow-up recommendations.
Intervention across the Diffusion Curve
How widely a screening test is adopted at a given time has implications for approaches to intervention. Interventions may be more effective if strategies are commensurate with a test's diffusion status. For example, in the early stages of test adoption, when baseline screening rates are modest, low-intensity strategies that have a wide reach (e.g., media campaigns) can be effective at increasing consumer and provider awareness. Although those who are quick to adopt the behavior (the early adopters) will likely respond to such general approaches, a substantial majority of the population will require more individualized efforts. Intervention then needs to take on more specificity to address the barriers to screening adherence. The most effective strategy will depend on the target audience and resources available, as well as how a screening program is organized (e.g., whether a country has a centralized screening program, or whether persons are members of an organized health care system, such as a health maintenance organization [HMO]).
As tests become more widely adopted, the relative importance of some barriers to screening will change, whereas others may persist or become more significant. For example, a lack of awareness among patients and providers that screening is indicated may be an initial impediment to screening. However, as the population's awareness increases over time, other issues such as access will emerge as important barriers to adherence. As people actually experience screening, other barriers may arise. For example, women's reports of painful mammograms may be important determinants of whether they obtain their next mammogram on schedule.51, 52 Not having the means to pay for screening (e.g., not having health insurance) or lacking a recommendation from a provider are likely to remain important barriers for some segments of the population, especially those who must pay for tests out-of-pocket or those without a regular health care provider.51
Early studies of patients and providers provided strikingly different explanations of the reasons for underscreening.53 Women tended to say they had not been screened because their physician had not recommended it or tests were not needed in the absence of symptoms,22 but physicians attributed the underuse of screening to such factors as cost.54 Self-reported data from the 2000 National Health Interview Survey demonstrated that not having undergone a recent colorectal examination was attributable most often to lack of awareness of the need for the test. In contrast, physicians responding to a provider survey of colorectal cancer screening practices55 stated that a lack of patient motivation was the most common reason for the lack of adherence to colorectal cancer screening.
Clearly, we have more to learn with regard to the actual barriers to screening, including the interplay between structural and individual factors and what it will take to overcome these barriers. It also is important that people receive the clearest possible messages about screening from their providers. Similarly, providers should be informed about patient's real barriers, which may be different from those assumed by physicians. It is critical to choose intervention approaches that are well matched to the population need and available resources at a given time. Even generally effective interventions (e.g., reminder systems) will not work if they do not address the prevalent barriers to screening (e.g., access). Therefore, the level of diffusion of a given screening test will influence the type of intervention strategy needed to promote change.
Successful strategies to enhance adherence to cancer screening have been directed toward multiple levels, including individuals, health care providers, health care and other organizations (e.g., work sites, churches, etc.), and communities. Strategies also have been organized based on the main barrier they address (e.g., lack of knowledge or negative attitudes, lack of access to services, or provider or system barriers to service delivery [available from URL: http://www.thecommunityguide.org/cancer (accessed 7/7/2004)]). For example, interventions to improve access to and therefore the use of mammography have included mobile mammography vans56 and free or low-cost vouchers.57 Such strategies have been particularly important in reaching women who have historically underutilized screening services.2 Changes in health policy also have enhanced access to services and contributed to increases in screening prevalence by mandating insurance benefits to cover the costs of screening tests (e.g., Medicare coverage for cervical, breast, and colorectal cancer screening).58
Other methods to improve uptake of screening address knowledge or attitudinal barriers and have been implemented at an individual level. These methods include telephone counseling59, 60 to address barriers to screening and letters tailored to an individual's particular characteristics.61–63 Efforts to address knowledge and attitudinal barriers also have been implemented at a population level (e.g., community-wide education programs). Print and electronic media have complemented strategies to increase public awareness and knowledge regarding the benefits of early detection and to publicize the availability of screening programs. Interventions designed to facilitate changes in provider behavior have been particularly important. Such interventions often operate at the organizational level. Significant improvements in screening adherence and referral rates have resulted from computerized office reminder systems and prompts as well as staffing and procedural changes, such as having designated prevention clinics or employing nonphysician staff for screening.7 Although some single-intervention strategies have been effective, combined approaches are frequently more so.2, 3
To our knowledge, the literature is generally consistent in finding that recent screening is negatively correlated with older age, racial/ethnic minority status, less education, and lower income.46, 64, 65 Income and education appear to be more powerful predictors of recent screening than race,66, 67 but the greatest disparities in screening appear to be correlated with access to health care. Indeed, the strongest predictors of underutilizing screening are not having a usual source of health care and not having health insurance coverage.21, 46 Persons who have not used other preventive services (e.g., have not had a health maintenance visit) are more likely to underutilize screening.68 Recent immigration to the U.S. often is found to be related to access and has been demonstrated to be negatively associated with recent screening.46 The interpersonal dimensions of health care, such as satisfaction with the quality of the care received, can influence the use of cancer screening services as well.69
Population groups that underutilize screening, including persons who do not undergo screening tests according to guidelines and those who have never had a test, present the greatest challenges for intervention programs. The most intensive intervention efforts should be tailored toward individual barriers for which interventions are appropriate (e.g., lack of knowledge or access). However, some of these barriers have been difficult or beyond the level of influence for individual investigators or program planners (for example, improving financial access to health services). These barriers require broader policy or regulatory interventions, such as changes in the organization and distribution of health care and the allocation of resources to health programs at the federal, state, and local levels.
Even those individuals without access barriers may not necessarily undergo screening according to guidelines, and a lack of adherence to recommendations can have adverse consequences for persons who otherwise see value in screening.70 An example of how an apparent lack of barriers may not be decisive comes from a study of women enrolled in Blue Cross Blue Shield of North Carolina, in which only 49% of the women were undergoing regular mammograms.71 In moving the focus from initial to repeat screening or maintenance of screening, still other types of interventions may be required.
The multiple factors that influence decisions to engage in cancer screening (e.g., patient preferences, physician recommendations, costs, health policies, etc.) have expanded the scope of intervention research from a behavioral tradition focused mainly on individual persons to a multilevel, interdisciplinary field. Over the years, many models have been proposed to improve our understanding of the multiple pathways by which health behavior, including screening, is influenced. These models have been used to guide intervention approaches in a variety of settings and populations.72–74 Recently, Zapka et al.75 expanded a widely used organizational framework (Chronic Disease Model) that recognizes the many factors that influence health care processes, including screening. The framework illustrates how interventions to promote screening at one level, such as improving communication between patients and providers, are insufficient without complementary efforts that consider the environment in which those interactions take place, such as practice setting, health benefits, reimbursement, and professional group standards. However, to our knowledge, until recently, research concerning screening has largely focused on patient and provider factors. Parallel research regarding other health behaviors, such as population-level or societal-level investigations of screening utilization, has not received the same degree of attention.76 A good example demonstrating how both individual factors and the larger health care environment are associated with adherence comes from Phillips et al.67 Their study revealed that women were more likely to undergo screening mammography if they lived in an area with no shortage of primary care providers, mammography facilities with reminder systems, and a higher HMO market share. To our knowledge, this is one of only a few investigations to examine environmental factors in addition to individual correlates of screening behavior.
Many behavioral, community-based, and system-based interventions have proven effective for a number of screening sites. The level at which interventions that enhance screening adherence have been sustained beyond grant funding or have been incorporated into routine clinical practice in the U.S. has been disappointing, however. This may be related, at least in part, to having no centralized system of health care in the U.S. Moreover, to our knowledge, no organized system exists to disseminate proven interventions, and those who are charged with referring people to screening may have no financial motivation to do so. The sad reality is that few of the interventions found to be effective on the basis of the research reported herein have been disseminated on a larger scale. Furthermore, these interventions often have not been designed for broad dissemination.77 For example, interventions created as part of a research study may have a level of complexity and cost that does not allow most organizations to adopt them after the research period has ended. Therefore, as interventions are designed, consideration should be given with regard to how they will be disseminated and sustained if found to be successful.
Over the years, investigators have refined various strategies to promote cancer screening. Much has been learned regarding the demographic, psychosocial, and health system factors that influence screening uptake as well as how to design effective interventions. What have we learned from our experience with existing screening tests and how can this knowledge guide future intervention efforts? Numerous systematic reviews already exist that document the effects of interventions to enhance adherence to cancer screening tests (e.g., the Guide to Community Preventive Health Services78 of the Centers for Disease Control and Prevention [CDC]; Cochrane Reviews,5, 79 and meta-analyses),2–4, 6, 7 and it has not been our intention to duplicate them. Instead, our approach has been to identify the lessons learned from the existing body of evidence.
Cancer screening and follow-up is a process that involves repeated interactions among individual men and women, the health care system, and the broader environment. This supplement examines lessons learned from intervention-oriented and related research addressing one segment of that process—adherence to screening recommendations and follow-up of abnormal findings. Of equal importance, but not examined herein, is intervention research pertaining to diagnosis, treatment, and survivorship. The articles that follow address the lessons learned and suggest future directions for enhancing screening adherence.
This supplement includes the contributions of many leaders in the field of cancer screening. A steering committee, with representatives from the National Cancer Institute, the American Cancer Society, the CDC, and academia, developed an outline for the supplement and invited authors to write articles. Early in the process of drafting the articles, the authors participated in conference calls to identify gaps in the topics addressed and to facilitate consistency in the use of terminology and format across the articles. At a workshop convened in June 2003, the authors presented early drafts of their articles, which then were critiqued by invited experts. The articles then were revised and submitted for peer review. The final product reflects this substantive input. We hope that the supplement will stimulate others to reflect on this knowledge so that all can learn from our collective experience.
The articles in this supplement to Cancer offer lessons learned from conducting research to promote cancer screening by addressing the following topics.
Perspectives on Behavioral and Social Science Research on Cancer Screening.
What lessons can be extracted from the vast behavioral science literature regarding cancer screening, and in which areas do we lack sufficient information from which to draw lessons? This article provides a broad perspective regarding social and behavioral science and cancer control.
Some Methodologic Lessons Learned from Cancer Screening Studies.
Methodologic issues are important to consider when evaluating whether interventions have achieved their desired outcomes. This article presents the lessons learned from measuring screening behavior and discusses aspects of study design, execution, quality, and external validity.
Lessons Learned from Community-Based Cancer Screening Research.
This article addresses contributions from the substantial body of intervention research concerning cancer screening originating in and conducted in community settings.
Interventions for Patients, Providers, and Health Care Organizations.
Much of the research regarding improving cancer screening rates has occurred within health care settings. This article summarizes the lessons learned from interventions in the health care setting that have sought to overcome barriers related to provider behavior, patient participation, health systems, and public policy.
Interventions to Improve Follow-Up of Abnormal Findings in Cancer Screening.
Intervention efforts do not end when individuals are referred to or undergo screening tests. An important step in the process to reduce cancer morbidity and mortality is the follow-up of abnormal findings. This article presents a synthesis of research in this area.
Screening Provision: A Perspective from Countries Utilizing Organized Screening Programs.
This article discusses intervention research in Westernized countries other than the U.S. What lessons can we learn from screening programs across diverse populations and health care systems?
Informed Decision-Making: What is Its Role in Cancer Screening?
This article reviews the growing area of informed decision-making as it relates to cancer screening. Research in this emerging area is especially important.
Examining the Cost-Effectiveness of Screening Promotion Interventions.
This article considers the lessons learned and those yet to be learned relative to measuring the costs and effectiveness of screening promotion.
Disseminating Effective Cancer Screening Interventions.
Can we assume that evidence-based interventions will automatically diffuse into practice? This article summarizes what is known regarding the diffusion, replication, and dissemination of intervention strategies to increase cancer screening.
The Future of Research That Promotes Cancer Screening.
Drawing from the lessons of the preceding articles, this article presents a perspective concerning our experience in designing, implementing, evaluating, and disseminating interventions to promote cancer screening and follow-up and proposes areas of investigation for future research and practice.
The authors thank Stephen Taplin, M.D., for his thoughtful comments and acknowledge his suggestions for improving this article.