Reducing Cancer Burden in the Population: An Overview of Epidemiologic Evidence to Support Policies, Systems, and Environmental Changes

Abstract

“Reducing Cancer Burden in the Population: Epidemiologic Evidence to Support Policies, Systems, and Environmental Changes” is a compilation of 11 reviews addressing aspects of primary prevention of cancer (early life factors, vitamin D, and periodontal disease and the oral microbiome); primary and secondary prevention (in the transgender population); surveillance following secondary prevention (personalizing follow-up of patients diagnosed with an adenoma based on colorectal cancer risk); tertiary prevention (physical activity as an adjuvant to cancer treatment, measurement of patient-reported physical functioning in cancer treatment trials, and implementation of palliative care recommendations); and all 3 (changing epidemiology of oral cancers). Authors discuss policies, systems, and environment (PSE) changes that may be supported by epidemiologic evidence, for example, how existing public health and clinical policies and guidelines targeting noncancer outcomes may indirectly reduce cancer burden and how some cancer control policies could be refined to enhance effectiveness. Authors also discuss where systems and environment changes are necessary to ensure routine implementation of up-to-date, evidence-based policies and guidelines. Two other articles discuss infrastructure to support identification of existing and emerging cancer problems that could be reduced or avoided, including by PSE changes. These frameworks may guide impactful cancer research relevant to cancer centers’ catchment areas, as well as cancer control efforts in countries where noncommunicable diseases including cancer are on the rise. The weight of the evidence of these reviews supports opportunities for PSE changes and infrastructure that could reduce the cancer burden in populations.

INTRODUCTION

As part of the US cancer control effort, the National Comprehensive Cancer Control Program of the Centers for Disease Control and Prevention (CDC) supports US states, tribes, and territories in their development of comprehensive cancer control plans. According to the CDC, “Comprehensive cancer control is a strategic approach to preventing or minimizing the impact of cancer in communities” (1). The plans are tailored to the cancer problems experienced by the residents of those areas. The plans typically describe the cancer problems in their community and include goals (e.g., reduce the colorectal cancer mortality rate by a stated percentage), objectives (e.g., increase the proportion of eligible persons who are up-to-date on colorectal cancer screening), and strategies to achieve those objectives (e.g., providers offer evening and weekend appointments to increase access; providers offer annual fecal immunochemical testing for eligible persons who are not otherwise able or interested in colonoscopy as the screening tool). As a requirement for receipt of CDC funding for cancer plans, recipients assemble coalitions of stakeholders to implement the plans. The CDC along with the American Cancer Society recently promoted that cancer plans and coalitions identify and implement policy, systems, and environment (PSE) changes to prevent cancer and reduce its burden (2). To support these efforts, we had as an objective for this 2017 issue of Epidemiologic Reviews to identify factors for which the epidemiologic evidence is ripe for supporting such societal level strategies for change to reduce cancer burden. Thus, the theme of the 2017 volume is “Reducing Cancer Burden in the Population: Epidemiologic Evidence to Support Policies, Systems, and Environmental Changes.”

We were unsure whether cancer epidemiologists were familiar with the concept of PSE changes for cancer control. Thus, in the Call for Papers, we promoted the theme with classic and soon-to-be classic examples of how our field's contributions directly supported cancer control PSE changes. The classic example we gave was the following. Epidemiologic evidence coupled with laboratory data definitively supported that smoking is a cause of cancer. That conclusion ultimately led to primary prevention approaches for smoking in the population, such as development of smoking cessation strategies and programs; national guidelines to screen patients in primary care for smoking and to offer cessation strategies; taxation of tobacco products; and clean indoor air laws prohibiting cigarette smoking (details of that evidence and interventions have been reported previously (3)). The soon-to-be classic example we gave was as follows. Subsequent epidemiologic evidence sufficiently supported that smoking reduces survival and otherwise adversely affects survivorship in cancer survivors. That conclusion ultimately led to tertiary prevention guidelines for those with cancer (4).

In the Call for Papers, we asked authors to consider submitting a brief proposal for their review in advance of submission. We provided guidance to ensure that the reviews were thematically appropriate. As we received full submissions, we noted obvious contemporary areas that were missing and reached out to junior and senior researchers conducting research on these areas. Several expressed surprise that their cancer-relevant substantive area might be poised to inform PSE changes. However, after discussion with us, some enthusiastically submitted a review. In the Call for Papers, we asked authors to do more than summarize the epidemiologic literature; they had to have as the context for their review the generation of summary data to inform translation to reduce cancer burden. When we received a high-quality review lacking this perspective, we asked the authors to provide discussion on opportunities to inform PSE changes.

After thorough peer review by experts around the globe, we were excited to accept 11 reviews covering aspects across the cancer control continuum. We were pleased with the richness and depth of the topics that the authors contributed. We were also impressed with the thoughtfulness in identifying the gaps needed to inform PSE changes, which compelled us to accept narrative reviews that were deemed important by the reviewers. Our intent for this issue was a series of systematic reviews (6 (5–10) of the 11 are) but, as readers can see, 3 of the articles are narrative reviews (11–13) and 2 include systematic reviews of already published meta-analyses (14, 15). The narrative review approach tended to be used when a topic was considered to be promising, but research findings were too inconclusive for new PSE action. We should disclose that reviewers did not always agree on whether narrative reviews, which tend to include authors’ perspectives and opinions, are appropriate now that systematic reviews are the norm. In addition, we realize that the structures of the narrative reviews included in this issue are not standardized and quite different from one another, making it somewhat challenging for readers. Nevertheless, we see great value in the included narrative reviews for informing PSE strategies for cancer prevention and control now or in the future.

Of the 11 reviews, 3 address aspects of primary prevention (10, 12, 14), 1 discusses both primary and secondary prevention (7), 1 covers surveillance following secondary prevention (5), 1 reports on primary, secondary, and tertiary prevention (6), and 3 address tertiary prevention (9, 13, 15). The other 2 reviews provide frameworks for infrastructure to identify cancer problems (8, 11); developing such infrastructure can be promoted through PSE changes. In the Call for Papers, we encouraged authors to submit reviews with global and life-course perspectives and reviews with focus on special populations. With respect to the global setting, 1 review provides a framework for infrastructure to identify cancer problems in countries shifting from predominantly communicable to noncommunicable diseases (11). With respect to life-course perspectives, 1 review addresses the time period from in utero through adolescence (12). With respect to special populations, 1 addresses the understudied and often underserved transgender population (7). Summarized here are the contexts for and some of the key points from these exceptional reviews.

PRIMARY PREVENTION

Vitamin D has received research attention as a possible cancer prevention agent for several decades. Within the last 10 years, major consortia such as the Cohort Consortium (16–23) have sought to provide estimates of circulating levels of 25-hydroxyvitamin D (25(OH)D) that may be cancer protective. Given the likely translational readiness of this evidence base for some cancer sites, we solicited a review from Dr. Alison Mondul from the University of Michigan and her colleague Dr. Demetrius Albanes from the National Cancer Institute (NCI), who has dedicated his career to identifying in observational epidemiologic studies and testing in randomized trials nutrients in the prevention of cancer. Mondul et al. (14) in “Vitamin D and Cancer Risk and Mortality: State of the Science, Gaps, and Challenges” summarize that higher circulating levels of the marker of vitamin D sufficiency (25(OH)D) are associated with a lower risk of colorectal and bladder cancers. They cover existing policies for vitamin D sufficiency for noncancer reasons, for example, rickets (24) and fall prevention in those over 65 years (25). Although screening for vitamin D insufficiency was widely promoted in popular media, as of November 2014, the US Preventive Services Task Force concluded that “the current evidence is insufficient to assess the balance of benefits and harms of screening for vitamin D deficiency in asymptomatic adults” (26). Although the policies for daily vitamin D intake could be indirectly beneficial for some, but not all cancers, the authors indicate gaps in the literature that make cancer-specific policies for vitamin D intervention, including screening for low levels, premature.

A very exciting, contemporary line of epidemiologic inquiry is the role of the microbiome in the etiology of cancer. About 10 years ago, Dr. Dominique Michaud and her Harvard colleagues published intriguing findings from a prospective study on periodontal disease, which is in part caused by oral pathogens, and an increased risk of pancreatic and other cancers (27, 28), findings that they expanded on in 2016 (29). Their work followed that of others but dealt carefully with confounding by smoking (an important cause of periodontal disease and a risk factor for pancreatic and many other cancers). Dr. Michaud, working with colleagues from the European Prospective Investigation Into Cancer and Nutrition (EPIC), subsequently identified that a higher level of circulating antibodies against Porphyromonas gingivalis, a bacterium involved in periodontitis, was associated with a higher subsequent risk of pancreatic cancer, while a higher level of circulating antibodies against commensal oral bacteria was associated with a lower risk (30). Other investigators have now directly investigated the oral microbiome and pancreatic cancer risk, and they have identified the same agent as being associated with increased pancreatic cancer risk in a prospective study (31). Given her leadership in this area, we solicited a review from Dr. Michaud at the Tufts University Cancer Center. In the review, “Periodontal Disease, Tooth Loss, and Cancer Risk,” Michaud et al. (10) provide a comprehensive overview of the mounting evidence base supporting periodontal disease—gingivitis and periodontitis—as at least a marker, if not a causal factor, in the development of several cancers, including pancreatic, lung, and head and neck cancers. Because this area is flourishing, several reviews have already been published. We were supportive of the authors updating the findings of those reviews with publications from the last 5 years. While Michaud et al. build the case for the importance of periodontal disease as a cancer risk factor because of its high prevalence, they point to information gaps before PSE changes are warranted. To fill information gaps, they call for improved measurements of periodontal disease for use in large epidemiologic studies of other cancers, including those that are less common, or to be able to study cancer in lifelong never smokers. They recommend that investigators use standardized protocols for the measurement of periodontal disease and microbiota so that data can be pooled across studies, which would increase the ability to study subpopulations or with strata of other exposures (e.g., never smokers). Even without knowledge of a clear mechanism, the authors describe the possibility of using periodontal disease as a risk stratification tool for cancer screening, for example, for targeting smokers with severe periodontal disease for lung cancer screening. Michaud et al. make the point that policies to increase access to dental care, including the treatment of periodontal disease, may indirectly benefit cancer risk.

Clarke and Joshu (12) from the Johns Hopkins Bloomberg School of Public Health were responsive to our welcoming of pieces addressing the life course. In “Early Life Exposures and Adult Cancer Risk,” they address early life factors covering gestation through adolescence that could influence, directly or indirectly, adulthood cancer risk. Cancer epidemiology has been so long focused on exposures in mid- and older life for research feasibility. Yet, it is entirely plausible that avoidance of early life cancer risk factors could reduce the burden of cancer in the population. The authors focus strategically on tobacco use, obesity, diet, and physical inactivity, known cancer risk factors that were identified in studies of individuals in mid- and older age. Clarke and Joshu highlight that most research on early life factors and cancer risk is mostly limited to breast cancer, and they point to huge gaps in the literature on this part of the life course for other cancers. They describe the potential biological actions of these early life exposures that may directly affect carcinogenesis during a particularly susceptible developmental/maturation period. They also suggest that exposures to risk factors early in life can hasten the time to reaching a sufficient cumulative exposure to a carcinogen. In their call for research on early life exposures and adult cancer risk, they recommend frameworks and tools that may help investigators conduct such research without having to rely solely on recalled exposures. With respect to policies, they point to those already in place for noncancer maternal and child reasons (e.g., recommendations against smoking and excessive weight gain during pregnancy) that are already indirectly benefiting susceptibility to cancer. They conclude that the cancer-specific policies for early life factors other than tobacco are not yet indicated.

PRIMARY AND SECONDARY PREVENTION

Over the past few years, advocates and the media have made society much more aware of persons whose gender identity or expression differs from that originally assigned to them at birth. Members of this special population tend to be medically underserved, and their chronic disease-related needs are understudied. Thus, we were pleased to receive the timely review, “Cancer in Transgender People: Evidence and Methodological Considerations.” In it, Braun, et al. (7) from Emory University describe factors that, in particular, may increase cancer risk in transgender persons, including the recognized higher prevalence of known risk factors such as smoking and alcohol use, obesity, physical inactivity, and human papillomavirus (HPV), as well as reduced access to and uptake of cancer screening. The authors point to the need for systematic collection of gender identity in state cancer registries, including those that are part of the NCI's Surveillance, Epidemiology, and End Results Program (32), to be able to quantify cancer incidence and mortality and to determine organ sites that may be of specific importance for subgroups of this special population. In addition, the authors call for empirical data from large prospective epidemiologic studies among transgender persons to identify population-specific risk factors, including possibly those related to use of gender-affirming sex steroid hormones and antihormones and social determinants, such as lack of health insurance, lack of provider understanding of trans health, and discrimination. Doing so, they argue, will allow us to be able to best intervene on cancer risk in this understudied and often underserved group.

SURVEILLANCE AFTER SECONDARY PREVENTION

One of the great cancer control successes we have had in the United States is the uptake of colorectal cancer screening and the associated decline in colorectal cancer incidence (3.2% per year over 2004–2013) and mortality (2.7% each year over 2004–2013) (33). The prevalence of colorectal cancer screening consistent with the 2008 US Preventive Services Task Force recommendations that were updated in 2016 (34) has increased substantially over the last 15 years, including a doubling of the prevalence in minority populations. In 2015, the prevalence of colorectal cancer screening was 66% in non-Hispanic whites, 60% in non-Hispanic blacks, 52% in non-Hispanic Asians, and 47% in Hispanic adults aged 50–75 years (35). We still have a way to go to reach the Healthy People 2020 target of 70.5% (36) and even farther to achieve the National Colorectal Cancer Roundtable's goal of 80% by 2018 (37). With an increase in the number of people being screened, including those who have never been screened and who may be enriched with colorectal cancer or precursor adenomatous polyps, the number of people who may require follow-up surveillance after the removal of a polyp will rise. We all want to dedicate limited health-care resources to provide the best care for patients to reduce cancer burden in the population, including those who are found to have a precursor polyp on screening. McFerran et al. (5) from Queen's University Belfast (Northern Ireland), University of Dublin Trinity College (Dublin, Ireland), and Memorial Sloan Kettering (New York, New York) contributed a review, entitled “Evaluation of the Effectiveness and Cost-Effectiveness of Personalized Surveillance After Colorectal Adenomatous Polypectomy.” Northern Ireland's Bowel Cancer Screening Program, launched in 2010, uses every other year fecal occult blood testing of men and women who are 60–74 years of age, which is followed up by colonoscopy for a positive screen when medically appropriate (38). In the United States, the US Preventive Services Task Force recommends colorectal cancer screening for men and women aged 50–75 years, but it does not indicate a preferred screening tool for those at average risk (34). In both settings, those who have an adenomatous polyp detected are surveilled subsequently. The authors performed a systematic review to address personalized surveillance postpolypectomy. By “personalized,” the authors mean targeting the surveillance strategy based on the likelihood of risk of colorectal cancer. In their review, they note that the literature supports that colonoscopy tends to identify people with small, low-risk adenomas, subjecting them to intensive surveillance and overmedicalizing them without clear benefit of reducing their risk of developing an advanced adenoma or frank colorectal cancer, wasting limited resources. They call for less intensive surveillance of individuals with low-risk polyps, which could include fecal blood testing and recommendation for risk factor modification. In other words, these personalized surveillance strategies could be implemented by policy (i.e., clinical practice guidelines) and systems (i.e., electronic medical records and referral systems) changes.

PRIMARY, SECONDARY, AND TERTIARY PREVENTION

In 2007, the International Agency for Research on Cancer concluded, “There is sufficient evidence in humans for the carcinogenicity of HPV 16 in the cervix, vulva (basaloid and warty tumors), vagina, penis (basaloid and warty tumors), anus, oral cavity and oropharynx” (39, p. 476) but that the evidence is limited for HPV 16 in the larynx, HPV 18 in the oral cavity and larynx, and HPV 6 and 11 in the larynx (squamous). A flurry of research on HPV-positive oropharyngeal cancer, its risk factors, and prevention followed. Since then, it has become clearer that HPV is a major cause of some cancers of the oropharynx, but less so of the oral cavity (40). In our view, the causal evidence is ripe for PSE changes for the prevention of oropharyngeal cancer and, thus, we were pleased to receive the review, entitled “The Health System and Policy Implications of Changing Epidemiology for Oral Cavity and Oropharyngeal Cancers in the United States From 1995 to 2016.” In their review, LeHew et al. (6) from the University of Illinois at the Chicago School of Dentistry discuss the need to separate oropharyngeal cancer from the general category of oral cancer to inform prevention, early detection, and cancer management strategies, including whom to target for these strategies and which providers should be engaged in preventive care and education. They remind us that some oropharyngeal cancers have HPV as a cause, while many oral cancers are caused by heavy smoking and alcohol drinking; that oropharyngeal cancer rates are increasing but survival tends to be good, while oral cancer rates are decreasing but survival tends to be poor; and that the populations burdened differ, the former tending to be younger, white men, while the latter tending to be older, black men. They point to the facts that access to oral health care is not uniform across populations and that education and lay understanding about oral and oropharyngeal cancers are limited and imperfect. With respect to PSE changes, they call for workforce development to serve medically and dentally underserved populations who tend to be at higher risk for oral cancer and for improved access to better-informed, organized, and coordinated medical and dental care.

TERTIARY PREVENTION

Patient-reported outcomes (PROs) are those directly reported by the patient and not interpreted by others rather than those assessed by the providers and investigators; many users cite the US Food and Drug Administration's PRO definition (41). PROs are measurements complementary to those assessed by providers/investigators in determining how well a treatment works and the potential adverse effects a treatment may cause. Examples of PROs include symptoms and quality of life and the topic of this review, physical functioning. Atkinson et al. (9) from Memorial Sloan Kettering in their review, “Patient-Reported Physical Function Measures in Cancer Clinical Trials,” posit that PRO physical functioning should be routinely measured in cancer clinical trials to be able to determine the likelihood that patient functioning returns to levels before treatment such that patients can resume their independence. They report that only about 1% of published cancer trials included such measures. They found that about two-thirds of these used 2 measures—the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire (42) and the Medical Outcomes Study Short Form 36 (43). The authors indicated that, while both of these tools are consistent with the Food and Drug Administration's PRO guidance (refer to Table 2 in “Guidance for Industry Patient-Reported Outcome Measures: Use in Medical Product Development to Support Labeling Claims” (41)), neither measure involved patient input into their development. The authors point to a physical function tool called the Patient-Reported Outcomes Measurement Information System Physical Function (PROMIS-PF) (44). This tool, the authors point out, was developed with the input of patients (albeit not solely cancer patients), which they feel increases the validity of the measurement. They conclude that valid and reliable tools for patients to self-rate their physical functioning at each study visit in cancer clinical trials and that can be used in treatment decision-making could lead to better outcomes in cancer patients, including the increased likelihood that patients maintain their independence. Routinely including such tools in trials would require policy changes and refinements, as well as systems changes (e.g., embedding tools assessing functional status into existing systems for collecting PROs).

Of the accepted reviews, the one by Atkinson et al. had the strongest critique, specifically the focus on patient self-rated physical function and not more broadly on patient self-reported physical health for use in cancer clinical trials. Atkinson et al. acknowledge the long history of use of PRO tools assessing physical health, such as the Functional Assessment of Cancer Therapy-General (FACT-G) physical well-being subscale, which was developed specifically for use in cancer treatment trials (45), and recognize that some of these tools are correlated with direct measures of physical functioning. The authors argue strongly that direct measurements of physical functioning not indirect correlates are also needed. We concurred with the authors’ perspective.

In 2011, the Institute of Medicine convened a workshop on the influence of obesity in cancer survival and recurrence. A report from the workshop indicated that several studies support that patients with early stage breast, prostate, and colorectal cancers who participate in physical activity after their diagnosis have better cancer-specific survival (46). Outcomes of the workshop included the identification of gaps in the literature on physical activity (as a facilitator of energy balance) in persons with cancer and recommendations for more research (46). Thus, we were delighted to receive a review from Dr. Kathryn Schmitz, a noted expert on exercise and cancer survivorship, who is now at the Penn State Health Milton S. Hershey Medical Center, and her colleagues, that provides an update beyond the evidence that was in existence at the time of the workshop. Led by Dr. Prue Cormie (15) from the Australian Catholic University, their review, entitled “The Impact of Exercise on Cancer Mortality, Recurrence, and Treatment-Related Adverse Effects,” provides a comprehensive and systematic evidence update. The authors report that the literature continues to support an exercise-associated survival benefit for cancer. Given the heterogeneity of the cancers, the nature of the physical activity interventions, and the designs of studies included, we were supportive of the review being a systematic review without additional meta-analysis. The authors do provide in the Discussion ranges of estimates of association or effect from the already published meta-analyses. Given these findings, the authors advocate for physical activity as an adjuvant therapy for cancer, not just exercise as a good thing that people with cancer might consider engaging in. To be able to use physical activity as a complementary treatment, the authors point to gaps in knowledge. They call for trials specific to a given cancer type testing specific physical activity interventions to inform the recommended amount of physical activity and the timing and duration of physical activity relative to the diagnosis date. The authors also report that cancer patients who are physically active experience fewer adverse effects and are less likely to experience severe adverse treatment effects. In the modern era of cancer care and treatment, guideline makers need to be aware that objective outcomes (e.g., cancer recurrence, cancer-specific death) need to be complemented by what patients report they are experiencing and what patients most care about, such as whether they have pain or change in their ability to taste or smell. The findings of this review are important, because they support that physical activity beneficially affects outcomes important to cancer patients. The literature on physical activity and treatment adverse effects is now large and includes both observational studies and trials. We were supportive of the approach the authors used in their systematic review: They present individual studies published in the past 5 years along with summary findings from earlier meta-analyses. Using this approach, readers can see how knowledge has evolved.

After noting that none of the submitted reviews covered palliative care for cancer patients, an essential cancer control strategy, we invited Dr. Thomas Smith, a leading clinical and research expert in this area at the Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins, to submit a review. In the United States, the implementation of palliative care programs and the training of palliative care physicians and other health-care providers has been slow, possibly, in part, because Americans view a recommendation of palliative care to mean that physicians have given up on them. This view is slowly changing as the evidence to support the recommendation of palliative care early after a cancer diagnosis has brought benefits to patients in pain control and longer life with better quality, and from a health system perspective, lower costs. To provide up-to-date, evidence-based recommendations, the American Society of Clinical Oncology recently released their Clinical Practice Guideline on “Integration of Palliative Care Into Standard Oncology Care” (47). The guidelines recommend that cancer patients diagnosed with advanced stage disease, irrespective of inpatient or outpatient status, should receive referrals for palliative care delivered by multidisciplinary teams at the same time that they receive treatment for their cancer. The guidelines also recommend the training of a dedicated workforce (palliative care specialists), as well as the training of oncologists in the provision of palliative care. Dr. Smith is an author of those guidelines.

The review contributed by Isenberg et al. (13), entitled “Implementing Evidence-Based Palliative Care Programs and Policy for Cancer Patients: Epidemiologic and Policy Implications of the 2016 American Society of Clinical Oncology Clinical Practice Guideline Update,” discusses what is needed to truly implement these guidelines and points to additional research needed. Isenberg et al. say that stakeholders across all parts of the health-care system need to support the implementation of palliative care, which means that policies, systems, and the environment must change. They argue that a key needed change to ensure widespread implementation of palliative care in cancer care is in the health-care funding model. The good news, they say, is that the funding model is changing now because of the Centers for Medicare & Medicaid Services’ expectation of value-based programs for care (48) and forces promoting care coordination (49), including for cancer (50, 51).

PSE CHANGES AND THE DEVELOPMENT OF INFRASTRUCTURE

Readers may notice that 2 reviews fall outside of the stated scope of the Call for Papers. On reading these submissions, we immediately recognized their value for the 2017 issue. It is equally critical to provide information needed to generate infrastructure to support the identification of cancer problems in places like the United States and in places making the transition from communicable to noncommunicable diseases as it is to provide evidence to support cancer control PSE changes. Here's why: Just because we have epidemiologic evidence to support a PSE change does not mean that such a change is needed in some populations, if the cancer-relevant issue is not a problem in a particular locale. Further, just because a cancer problem may exist in a certain locale does not mean that that a locale or institutions located in that locale are organized to conduct research to provide evidence to support intervening on that cancer problem. Developing such infrastructure involves PSE changes.

With respect to the United States, the review contributed by Tai and Hiatt (8) from the University of California, San Francisco, entitled “The Population Burden of Cancer: Research Driven by the Catchment Area of a Cancer Center,” is highly relevant to NCI's current requirement that its designated cancer centers define their catchment area, identify the cancer problems in their catchment area, and conduct research relevant to those problems. Many cancer centers remain confused about how best to meet NCI's requirement. In addition, inexplicably during site visits, reviewers of cancer centers (who themselves are members of other cancer centers) often apply their own unwritten rules to how they think the catchment area should have been defined and to what constitutes research relevant to the catchment area, which is patently unfair. Tai and Hiatt give examples from the literature and from their comprehensive cancer center to help illuminate this process. Cancer centers will now be able to cite this review in support of their approach or at least in support of no right or wrong approach to defining the catchment area and its cancer-related problems. Of course, those citing this review need to keep in mind the reason why NCI has this requirement of recipients of Cancer Center Support Grants (52): to actually make a difference in the cancer burden for those living in the communities where cancer centers are located. Keeping in mind this reason means that cancer centers must be purposeful in their recruitment of members with appropriate cancer expertise and in their organization of research that truly has the promise to inform strategies, including PSE changes, to address cancer-relevant problems in the catchment area.

In the United States, we take for granted the established, but complex process for the collection and curation of high-quality cancer incidence and mortality data. In some US states, reporting of cancer diagnoses to the state cancer registry is mandated by law (e.g., Maryland (53)). In other words, policy and systems changes were needed along with funding and infrastructure support from the CDC (54), as well as a professional organization that develops standards for cancer registries, trains registrars, and certifies cancer registries (55) and also a documented need for such data by cancer control planners among other needed factors to consistently yield such data. The review by Piñeros et al. (11) from the International Agency for Research on Cancer, entitled “A Global Cancer Surveillance Framework Within Noncommunicable Disease Surveillance: Making the Case for Population-Based Cancer Registries,” reminds us that one of the most pressing issues as countries transition from communicable diseases to noncommunicable diseases predominating is the inability to accurately capture cancer incidence and mortality. Without such data, countries and organizations will not know which context-specific PSE changes should be implemented to be able to prevent or reduce an emerging cancer problem. The authors call for country-level surveillance designed specifically to capture cancer events and to provide a framework that countries may use to monitor the progress of cancer control. Perhaps noncommunicable diseases researchers in global settings could use this article to help explain to their ministries of health why resources should be dedicated to developing and maintaining a full-fledged cancer registry, rather than simply hospital and clinic cancer registries. Those latter registries tend to undercount the total number of cases in an area, give only a partial view of the stage distribution at diagnosis, and are not mapped to population denominator needs to calculate cancer rates. Most US-based cancer researchers and cancer control planners rely on cancer registry data, whether they realize it or not. They too will learn from this review what it takes to have a proper population-based cancer registry.

MUSINGS

As epidemiologists, we have a role to play in translation to reduce the cancer burden in populations, which brings me to an observation we made about the 2017 issue. The Call for Papers produced fewer inquiries about submissions, presubmission abstracts for early editorial input, and full manuscript submission than previous recent issues. Although it is possible that we did not adequately distribute the Call for Papers or we were not sufficiently clear about what we intended for the issue's reviews, I am concerned about other possible explanations related to translation. Do we cancer epidemiologists not routinely seek opportunities for translation to inform the research we choose to conduct and as the natural next step after drawing inferences from our research? Does our field not recognize our individual and collective responsibility as epidemiologists to inform translation? These 2 possible explanations are related and imply that we are oblivious and/or we are shirking our responsibility and roots as epidemiologists. Translational epidemiology is the heart of our discipline, and over the past 10 years, the field has had a resurgence in the call for translation to make a public health difference (56). Are cancer epidemiologists not familiar with the concept of PSE change as a cancer control strategy? I hope the 2017 issue can familiarize epidemiologists and other interested public health and clinical professionals with the PSE change strategy. Are we so engaged in our primary research that we do not have the bandwidth to prepare such a review that involves the extra challenge of providing PSE changes as the context for the reviews? This possible explanation is related to the prior one in that potential authors would have had to learn about what PSE change is and its role in cancer control, which they may not have had the time to do. Many academic epidemiologists feel overworked and overburdened. We together need to figure out strategies that will allow us the time to fully participate in knowledge synthesis and transfer.

Have we not identified many “new” cancer risk and prognostic factors for which evidence supports PSE strategies for prevention and intervention now? This one implies we are not being creative in our thinking. Dr. Graham Colditz from Washington University, who served on the review committee for this issue, and his colleagues recommended that the field implement better what we already know about intervening on cancer now, given that we know the cause of 50% of cancer in the population—primarily tobacco, obesity, and physical inactivity (57). However, they did not say that nothing is left to be discovered or that epidemiologists should cease searching for new risk and protective factors in those at risk for cancer and in those with cancer at risk for poor outcomes in general and importantly in populations disproportionately burdened with cancer. They just wanted us, as a society, to act now on the known risk factors instead of us, as epidemiologists, restudying the same risk factors over and over with just slight nuances, if any.

Do we as epidemiologists lack an appreciation of when the literature is ripe for informing PSE changes? That's quite possible. We are trained to root out bias and confounding and to want large relative risks and narrow confidence intervals. We want lots of data, including in strata. When we do engage in knowledge synthesis of observational studies, we report a summary estimate but are always cautious—association, not causation. For this step, we perhaps can learn from interacting with policy makers. They typically must act without complete information. If we epidemiologists understood better the extent and level of evidence policy makers need in order to take action, perhaps we would be less conservative in our conclusions, would be explicit in describing in detail the next steps for translation (rather than the usual “more research is needed”), and would reach out to policy makers to ask how we can help as translational cancer epidemiologists.

A WAY FORWARD

I hope that readers find that the 2017 issue of Epidemiologic Reviews has met our objective of providing epidemiologic evidence bases and paths forward for generating additional epidemiologic evidence bases to support PSE changes to reduce cancer burden in the population. Although outside of the Call for Papers, the reviews on developing infrastructure to support the identification of existing and emerging cancer problems that could be reduced or avoided including by PSE changes may be helpful for guiding impactful catchment area relevant research and for guiding cancer control efforts in countries where noncommunicable diseases, including cancer, are on the rise.

After reviewing the evidence, the authors were thoughtful in their discussion of PSE changes that could be supported by the evidence. These reviews pointed to existing policies (e.g., national and professional society guidelines and recommendations) that already may be, albeit indirectly, reducing the burden of cancer in the population, including those aimed at avoiding insufficient vitamin D intake and avoiding maternal excess weight gain and smoking during pregnancy (possible benefit to offspring later in life). However, with improved, supportive evidence bases, these policies could be expanded and tailored to cancer. Systems (e.g., electronic medical records systems) would need to be modified to accommodate recommendations that are even more complex than they currently are. Two of the reviews pointed to increased dental health-care access as a PSE change for cancer control, specifically in mitigating cancer burden associated with oral cancer, especially in black men, and possibly periodontitis-associated cancers (e.g., pancreas, lung). The epidemiologic evidence base for the former is well established, while the latter is building rapidly. Dentists have been involved in the cancer control effort to reduce the burden of smoking- and alcohol-associated oral cancers by detecting lesions early, although evidence is lacking to support a reduction in death from oral cancers in persons screened (58). Given that oropharyngeal cancers are difficult to detect, but oral HPV is detectable in oral rinses (59), I can also envision the possibility of dentists leading the effort to detect persistent oral HPV, if the tools to validly and reliably detect the virus become available and persistent oral HPV is shown to be associated with increased risk of HPV-positive oropharyngeal cancer. If the evidence base continues to support the periodontitis-cancer link, the review points out that dentists might also play an expanded role in cancer control beyond treating periodontal disease by identifying patients who might be targeted for cancer screening (e.g., lung) or other nondental interventions.

The reviews also pointed to existing policies that directly reduce the burden of cancer in the population that could be refined for more effective cancer control, including by personalizing colorectal cancer surveillance following adenomatous polyp removal and by “prescribing” physical activity to cancer patients. For the former, well-established consensus clinical guidelines exist for subsequent surveillance by colonoscopy of average risk patients when they were initially screened based on their histology, size, and multiplicity (60). Perhaps these guidelines could be modified for the lowest risk adenomas by increasing the surveillance interval and/or switching technologies from colonoscopy to the fecal immunochemical test. For the latter, many groups already recommend exercise for cancer patients (61, 62). What is needed is deepening of the evidence base: What is the right dose, timing, and nature of activity based on the patient and the patient's cancer type and stage? Also needed are systems and environment changes for routine implementation given that cancer care providers may not be equipped to prescribe and support their cancer patients in personalized physical activity. One of the reviews noted an example where a policy (i.e., professional society guideline) exists and is strongly worded, clear in intent, and based on an up-to-date evidence base (i.e., cancer palliative care). In that example, systems and environment changes are needed now for complete, routine implementation of cancer palliative care, including for workforce development.

One review called for the inclusion of direct measures of physical functioning in the conduct of cancer treatment trials. That view is consistent with the Food and Drug Administration's promotion of greater inclusion of PROs in cancer treatment trials, including physical functioning, to be able to characterize the adverse effects of these treatments on patients’ quality of life as balanced against survival. The Food and Drug Administration indicated that “International stakeholder collaboration and continued research into optimal practices for PRO and other clinical outcome assessments are necessary to advance a common framework for generating and reporting rigorous patient-centered data from cancer clinical trials” (63, p. 1553). On the basis of the “heated” back and forth between the authors and the reviewers of this review, it seems to me that more work is needed on the particulars of physical functioning that are the most important to patients and that are acceptable to both patients and those actually conducting the trials with respect to burden of data collection.

One review pointed to a policy change that is simple to describe but likely harder to implement without fierce demands from stakeholders, health-care providers, and academia; that is the collection of individual-level information on sex by using categories reflective of the distribution of gender identity by cancer registries. By cancer registries, I mean state cancer registries, including those that participate in the Surveillance, Epidemiology, and End Results Program, and the systems in place at hospitals, clinics, and pathology laboratories that feed new diagnoses into state cancer registries. Similar to other understudied populations, which usually equates to underserved populations, being able to address cancer problems specific to them and being able to tailor PSE changes to them require that cancer registry data be fully annotated with individual-level sociodemographic characteristics. Funders calling for research to study cancer problems and their causes and for research aiming to develop and evaluate interventions in the trans population may create a demand for the required collection of categories of a wide distribution of gender identity by cancer registry data.

In the Call for Papers, we said that the overall goal of this issue is to support the development and implementation of changes in policy (e.g., prevention guidelines, laws), systems (e.g., in primary care or cancer care, including precision prevention and precision oncology), and environment (e.g., in the physical environment in neighborhoods, schools, workplaces), that is, to support translation of epidemiologic research findings to reduce cancer burden in the population. Some of these PSE changes outlined above could happen now, and others need an improved or expanded evidence base. However, what we did not ask authors to explicitly comment on is who should be responsible for the translation step. Who should be responsible for ensuring that these evidence bases reach sufficiency? That's us cancer epidemiologists who currently work in these areas, as well as epidemiologists-in-training who seek the challenge of working on these not easy-to-study areas (e.g., early life exposure associations with adult cancer risk). We do this already: We identify gaps in the literature, and we or other epidemiologists attempt to fill them. However, I suspect that most of us are not systematic in mapping and then prioritizing evidence gaps that could inform particular PSE strategies as a way to drive our personal research agendas. Perhaps we should, at least at times, use this strategy to make translation more efficient. Who should modify the policies? As epidemiologists, we should participate in professional society guideline panels and independent bodies that advise federal, state, and local governments on cancer control along with those from other relevant disciplines. Many of our colleagues do participate, but we often find ourselves as the only voice from the population perspective. So, we need to speak more loudly and be sure to facilitate the participation of epidemiologists on future panels. We also have a role to play in educating policy makers both proactively (as appropriate) and reactively when invited to give input. We should participate fully in these activities while being mindful of potential real or perceived conflict-of-interest boundaries. Who should modify systems and environment? Here's where epidemiologists can partner with other stakeholders to help educate and guide those responsible for implementing and maintaining systems and environment. For example, participate as an academic expert in developing your state, tribe, or territory comprehensive cancer control plan and join your local or state cancer coalition as an individual or as a representative of your academic institution.

We did not intend for the 2017 issue to be comprehensive, so I am certain that so many other topics could have been reviewed for this issue. If this issue did not include a review on a topic that you think is ripe for PSE change for cancer control, perhaps you will now be motivated to prepare one with a translational epidemiology perspective!