Cancer Survivors in the United States: Prevalence across the Survivorship Trajectory and Implications for Care

This article is featured in Highlights of This Issue, p. 479

The number of people diagnosed with cancer during their lifetime has been steadily increasing. This trend is due to 2 complementary phenomena. The first is an improvement in cancer survival rates, driven by advances in early detection and improvements in cancer treatment (1). The second, and most significant, is the aging of the population (2). Because cancer incidence rates tend to increase with age, an aging population results in progressively more people being diagnosed with cancer. It is estimated that two-thirds of all cancer survivors will be age 65 or older by 2020 (3). The growing population of cancer survivors will put pressure on a healthcare system in which cancer drug shortages are increasingly common and the demand for oncology services is poised to outpace the supply of oncologists (4–6). The growing number of older survivors also presents a unique challenge to the healthcare system because older cancer survivors are more likely to have multiple chronic diseases and tend to experience poorer physical functioning than younger survivors (7, 8).

As the number of cancer survivors increases so does the cost of cancer care. By 2020, it is estimated that population growth alone will escalate annual costs of cancer care by 27% (9). Health care costs in the first year after a cancer diagnosis tend to be higher than annual costs thereafter for survivors who are not in their last year of life (9, 10). Nevertheless, among survivors who are more than 1-year postdiagnosis, annual healthcare expenditures are double that of the general population, suggesting that the economic burden of cancer in terms of medical expenditures is both considerable and persistent (11).

Cancer survivors' need for medical and supportive care services evolve from the point of diagnosis forward. In general, primary treatment ensues shortly after diagnosis with therapy typically consisting of surgery, radiation and/or chemotherapy. Cancer and its treatment can precipitate a range of physical and psychosocial sequelae that may require additional care; most commonly, functional limitations, fatigue, pain, cognitive changes, body image concerns, sexual dysfunction, psychosocial issues, and socioeconomic problems (12–16). The period immediately following the end of treatment presents another set of challenges as cancer survivors experience changes in the frequency of contact with their healthcare team, manage the lingering side effects of treatment, and resume important social roles and activities—all of which can precipitate feelings of distress (17–19). Following primary treatment, survivors' health service needs will vary depending on their clinical trajectory and treatment toxicities, which can vary substantially. Although some survivors will require no additional treatment others will undergo maintenance therapy (e.g., hormonal therapy) to lower their risk of recurrence. Survivors diagnosed with advanced or metastatic disease may receive ongoing treatment in an effort to slow or halt the progression of their disease or to palliate symptoms. In many cases, the acute effects of treatment begin to subside as time since diagnosis increases (15, 20, 21); however, a proportion of survivors will experience lingering side effects of their treatment that can persist for many years, negatively affecting quality of life (14–16, 21–23). Cancer treatment can also precipitate health problems that arise months or years after the end of active treatment. These late effects include second malignancies, disorders of the cardiovascular, genito-urinary or gastrointestinal systems, gonadal toxicity and endocrine disorders, compromised pulmonary function, and neurological and neurophysiological sequelae (24, 25). Data from the National Health Interview Survey, in which over 60% of cancer survivors were 5 or more years postdiagnosis, suggested that approximately 30% of survivors were in fair or poor health, 17% were unable to work due to health problems, and 58% had one or more functional limitation (26). Across each of these outcomes, cancer survivors had poorer health and functioning than people without a cancer history (26).

Survivors at the end of life report worse self-reported health, more activity limitations and lower values on the Health Activities and Limitations Index compared to other survivors (27). At the end of life, contact with the healthcare system becomes more frequent as survivors pursue salvage therapy, palliative care, or hospice. That over one quarter of total Medicare payments go to beneficiaries in their last year of life reflects the intensive medical care received during this period (28). In fact, healthcare costs for survivors in the last year of life are higher than costs incurred at other times following diagnosis (9).

To inform cancer survivorship research and health care practice and policy, this manuscript will describe the prevalent population of cancer survivors in terms of overall numbers and prevalence by cancer site and time since diagnosis. To estimate the extent to which the number of cancer survivors is estimated to increase over time, prevalence estimates will be projected through the year 2022.

Cancer prevalence was modeled using the Prevalence Incidence Approach Model (PIAMOD), which calculates prevalence from cancer survival and cancer incidence all-cause mortality (29, 30). The method first fits parametric models to incidence and survival data. Complete prevalence is then calculated and projected as the sum of the incidence and survival models product. In this study, we used the first malignant cancer by cancer site diagnosed between 1975 and 2007 from the 9 oldest registries in the Surveillance, Epidemiology, and End Results (SEER) program. The sample comprises children and adults who have been diagnosed with a broad range of cancers. However, cases of in situ disease and basal or squamous cell skin cancers were excluded from the sample. The most recent year of available data (2008) was excluded due to anticipated undercounts because of reporting delay. Incidence counts were estimated by applying SEER incidence proportions by cancer site, sex, year, age, and race to the respective U.S. populations. U.S. populations (1975–2009) and projections from 2010 through 2022 were obtained from the U.S. Census Bureau. The National Center for Health Statistics U.S. mortality data from 1969 to 2008 were obtained from the SEER*Stat software, and projected mortality rates for 2009 to 2022 were calculated by applying the 2006 to 2008 mortality rates by age and sex to the respective U.S. populations. Complete cancer prevalence by years since diagnosis, for years beyond 2007 (i.e., the last year of data) was projected assuming constant future mortality, incidence, and survival trends, but dynamic projections of the U.S. population. The modeled prevalence estimates and projections were adjusted to match the 1/1/2009 cancer prevalence reported in SEER Cancer Statistics Review, 1975 to 2009 (1) that uses the prevalence data more directly and with fewer assumptions. For each cancer site and sex, we calculated an adjustment by comparing the 1/1/2009 modeled prevalence with the 1/1/2009 observed prevalence (1). The cancer-specific adjustments were applied to the whole series of modeled prevalence from 1975 through 2022. The PIAMOD method also provides estimates by time since diagnosis in years. We have used 0 to less than 1 years from diagnosis to represent survivors receiving more intensive care in the first year after diagnosis, 1 to less than 5 years survivors in a more intensive monitoring phase, 5 to less than 10 years, 10 to less than 15 years, and 15+ years to represent long-term survivors in the survivorship trajectory.

The number of people currently living who have been diagnosed with cancer during their lifetime has been steadily increasing since the 1970s (Fig. 1). As of January 1, 2012, there were an estimated 13.7 million people in the United States who had been diagnosed with cancer (Table 1), which is approximately a 3.4 fold increase compared with 1975. Over the next decade, the number of cancer survivors is projected to increase by 31% to almost 18 million¹, which represents an increase of more than 4 million survivors in 10 years.

Women with breast cancer account for 22% of survivors, whereas men with prostate cancer account for 20% of survivors (Fig. 2). Although lung cancer is the second most common cancer diagnosed in men and women, only 3% of survivors were diagnosed with lung cancer. It is estimated that 64% of cancer survivors have survived 5 years or more; 40% have survived 10 years or more; and 15% have survived 20 years or more after their diagnosis (Fig. 3). Although the proportion of long-term survivors is quite high overall, there is considerable heterogeneity in the proportion of long-term survivors by cancer site (Fig. 4). For example, of the 245,022 cervical cancer survivors currently living in the United States, 83% have lived 5 years or more after diagnosis. In contrast, of the 412,226 lung cancer survivors currently living in the United States, only 42% have survived 5 years or more after diagnosis. Generally speaking, the proportion of long-term survivors for a particular cancer is both a function of the 5-year survival rates for that cancer as well as the median age at diagnosis such that the proportion of long-term survivors will be relatively high for cancers with a high 5-year survival rate but a low median age at diagnosis.

The number of new cancer cases and the number of long-term survivors is projected to increase over the next decade (Fig. 1). From 2012 to 2022, the number of survivors who are less than 5 years from diagnosis will increase from 4.9 million to 6.0 million or 22%; whereas the number of survivors who are 5 years or more from diagnosis will increase from 8.7 million to 11.9 million or 37%. Over the next decade, the relative proportion of survivors who are 5 years or more from diagnosis will increase from 64% to 67% with a corresponding decrease in the relative proportion of survivors who are less than 5 years from diagnosis. Of note, the largest relative increase in the number of cancer survivors by time since diagnosis will be among people who are 15+ years postdiagnosis where prevalence will increase from 3.4 million or 25% of all survivors in 2012 to 5.0 million or 28% of all survivors in 2022.

Over the next decade, prevalence across the survivorship trajectory is projected to increase as a result of increasing cancer incidence rates associated with the aging of the population and improvements in long-term survival rates. Women with breast cancer and men with prostate cancer represent the 2 largest groups of cancer survivors due to the relatively high incidence and low mortality rates associated with these diseases. Women are overrepresented among survivors who are 10+ years post diagnosis, because of the large proportion of women diagnosed with breast cancer, of which 90% will survive for 5 years or longer (31), as well as the longer life expectancy of women (32). Although prevalence is expected to increase across the survivorship trajectory, the largest group of survivors, both currently and moving forward, are those who are 5 years or more from diagnosis.

The burgeoning cohort of survivors where the absolute number of survivors is projected to increase by approximately 4.3 million survivors over the next 10 years presents a significant challenge to the health care system. This is especially true given that most survivors can now expect to live longer after diagnosis and often have complex needs stemming from chronic and late effects of treatment as well as comorbid diseases. The growth of the cancer survivor population will drive up the overall cost of cancer care an estimated 27% by the year 2020 from 2010 levels with more acute escalations in costs projected for certain subgroups of survivors (9). Costs for prostate and breast cancer survivors in the continuing phase of care (i.e., ≥1-year postdiagnosis but >1 year from death) are projected to increase 42% and 32%, respectively (9). Although annual costs for survivors in the continuing phase of care tend to be low compared with survivors who are either less than 1-year postdiagnosis or in the last year of life, the large number of survivors in the continuing phase of care is a significant driver of healthcare costs.

Over the next decade, a coordinated agenda for research and practice is needed to better understand and address the medical, psychosocial, and practical needs of cancer survivor across the survivorship trajectory. To this end, we propose the following recommendations.

Clinical guidelines for survivorship care have been issued by professional societies and the Institute of Medicine (33). However, to date, guidelines pertaining to survivorship care have been largely based on consensus rather than empirical data. The formal evaluation of follow-up care models is in its infancy, and most comparative studies in this area have focused on survivors at low risk for adverse health outcomes (33). Few studies have evaluated the cost effectiveness of different approaches (33). Risk-based follow-up care where the intensity of follow-up care is matched to survivors' risk for long-term and late-onset treatment effects is increasingly being explored. For example, risk-based follow-up care in the context of a shared care model has been proposed for the long-term follow up of childhood cancer survivors (34, 35). In addition, risk stratification of follow-up care is a central component of the United Kingdom's National Cancer Survivorship Initiative (36). However, as with other models of care, evaluation is needed to assess whether a risk-based model results in improved clinical, cost, quality of life, and satisfaction with care outcomes.

To develop and evaluate models of follow-up care, longitudinal data on current and emerging long-term and late effects associated with different cancer treatments are needed to characterize survivors' level of risk across a range of psychosocial and clinical risk factors (37). Extant clinical trials, epidemiology cohorts and consortia, data from integrated Health Maintenance Organizations, and population-based studies provide a wealth of data that can be leveraged for outcomes research with cancer survivors. A report summarizing the activities of the 2011 NCI-sponsored workshop, “Utilizing Data from Cancer Survivor Cohorts: Understanding the Current State of Knowledge and Developing Future Research Priorities,” highlights examples of how each of these data sources can be used for cancer survivorship research and recommends priority directions for future research, data collection considerations, and a summary of resources needed to advance survivorship research (38). A recognized challenge of using existing data for survivorship research is that research questions are constrained by the data available, and many existing cohort studies lack detailed treatment exposure data and data on tumor pathology. Thus, the need persists for a longitudinal cohort that captures both detailed information about the tumor and treatment history, biospecimens for future molecular, genetic, and genomic research, as well as patient-reported outcomes, characterizing survivors' health behaviors and their physical, psychosocial, and practical needs after the conclusion of cancer treatment.

Over the last decade, technological advances have given rise to electronic medical records (EMR), personal health records, real-time capture of biological and self-report data, and platforms to deliver interventions via mobile devices. Health IT in particular has the potential to improve care coordination by giving all members of the health care team access to a survivor's health information. Recent research found that 75% of cancer survivors felt it was very important for their health providers to have access their medical information electronically—suggesting a desire for communication and coordination among their providers (39). Unfortunately, many existing electronic record systems are proprietary and only accessible to providers practicing in a given healthcare setting (40). Lack of interoperability among different EMR systems is an important barrier to delivering appropriate long-term follow-up care to cancer survivors because providers outside of the cancer center (e.g., PCPs and other specialists) do not have direct access to cancer and treatment information and must rely on the survivors themselves to both possess and communicate a summary of their cancer and treatment. With the growing population of survivors, it is critical to enhance the capacity for health information exchange across health care settings. Current efforts to address this problem include the Health Information Technology for Economic and Clinical Health (HITECH) Act that was passed to incentivize the “meaningful use” of electronic health records and electronic health information exchange (41) and an alternative proposal by the President's Council of Advisors on Science and Technology who advocate creating a universal exchange language to enable the sharing of health information across electronic record systems (40). However, the impact of these approaches is yet unknown.

Advances in mHealth represent another vehicle for connecting survivors with extensive information about their disease and function as a resource to support survivors in managing their own health (42, 43). mHealth technology can also support effective symptom management by providing a way to monitor symptoms and other health problems and communicate this information to a healthcare provider (44, 45). As the use of mHealth technology increases, targeted outreach may be needed to overcome the “digital divide” and avoid social disparities in access to these promising new tools (46, 47). Further, because most cancer survivors are over the age of 65 (3), it is important to improve the acceptability and usability of mHealth technology among older adults. As of 2012, only 21% of older adults used a mobile device to access the Internet and 13% owned a smart phone (48). It is possible that a combination of functional limitations (e.g., limited manual dexterity, impaired vision) and low perceived utility of mobile devices accounts for the relatively low uptake among older adults. However, efforts are needed to fully characterize and address the range of factors affecting the use of mHealth technology among this population. Efforts to evaluate the reliability and validity of new technology for all segments of the population needs to keep pace with the development of novel tools, and evaluation methods are needed to provide timely data about this rapidly evolving field (49). Ongoing efforts and partnerships especially between academic researchers, information technology companies, and cancer care specialists in diverse communities are needed to ensure that the devices, new platforms, and associated applications are based on cutting-edge biomedical and behavioral research (49).

Research is needed to better understand the needs of adult survivors who are more than 5 years post diagnosis—a group representing 64% of all cancer survivors. Most of what is known about cancer survivorship has stemmed from research conducted among breast cancer survivors. For the field of cancer survivorship to move forward, it is critical to gain a better understanding of the needs of survivors with other diagnoses and across the posttreatment trajectory. Finally, there is little research on survivors with recurrent disease or second cancers. Survivors with recurrent disease report distress, lack of energy, difficulty sleeping, pain, worry, and sexual dysfunction (50–52). However, there are few studies and fewer interventions dedicated to this clinical population.

Research in palliative medicine has shown evidence that palliative care interventions are associated with lower costs of care, as well as increased quality of life and improved symptom management. Benefits of palliative care interventions also extend to grieving family members. In addition, recent studies provide preliminary data that early palliative care interventions increase not only quality of life, but also quantity of life (53, 54). Expanding research in this area to evaluate the optimal timing of palliative care interventions, concurrent palliative and oncologic treatments, and symptom management to those living significant periods of time with advance disease (e.g., Stage IV breast cancer) could lead to both improved patient experiences in the last year of life, as well as decreased overall health care costs to patients, payers, and health systems.

There were limitations to our analysis. Our estimates of cancer prevalence were based on cancer incidence and survival from the SEER-9 areas, which do not cover the entire United States. The SEER areas had lower incidence rates than most other states and have been found to have higher socioeconomic status, greater urban population, and more specialty care than the rest of the U.S. population. Our prevalence estimates differ somewhat from estimates used in the calculations of costs of cancer care reported elsewhere (9). In both studies, the prevalence projections were based on constant incidence and survival trends but dynamic population trends. Thus, they reflect the impact of population dynamics, especially the aging effect, on the overall burden of cancer, under currently disseminated cancer control policies and interventions. Different future assumptions of survival and incidence trends have shown to have a smaller impact on projections of the number of cancer survivors compared with the aging of the U.S. population (9). A study of costs of cancer care (18) shows that using constant future incidence and survival, we underestimate the number of cancer survivors for kidney and melanoma cancers, for which incidence has been increasing and slightly overestimate survival for cancer sites that have decreasing incidence in the most recent years, for example, cervix, ovary, breast, prostate, colorectal, and lung cancers.

The proportion of cancer survivors in the United States is projected to increase 31% by the year 2022. Owing to advances in early detection and cancer treatment, people are living longer after a cancer diagnosis, giving rise to a growing proportion of long-term survivors. A multipronged approach is essential to address the needs of cancer survivors across the survivorship trajectory. Efforts are needed to identify effective and efficient models for delivering long-term follow-up care; develop infrastructure to collect long-term clinical and patient-reported outcome data from survivors; harness health IT and other technologies that facilitate care coordination and improvement in survivors' long-term health outcomes; address important knowledge gaps about long-term survivors; and improve integrative palliative care. Progress in these areas is critical to achieving optimal clinical, cost, quality of life, and satisfaction with care outcomes.

No potential conflicts of interest were disclosed.

Conception and design: J.S. de Moor, A.B. Mariotto, C. Parry, C.M. Alfano, L.S. Padgett, E.E. Kent, L.P. Forsythe, J.H. Rowland

Development of methodology: A.B. Mariotto, C. Parry, C.M. Alfano, S. Scoppa, J.H. Rowland

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): J.S. de Moor, A.B. Mariotto, C.M. Alfano, E.E. Kent, S. Scoppa, M. Hachey, J.H. Rowland

Writing, review, and/or revision of the manuscript: J.S. de Moor, A.B. Mariotto, C. Parry, C.M. Alfano, L.S. Padgett, E.E. Kent, L.P. Forsythe, S. Scoppa, M. Hachey, J.H. Rowland

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): S. Scoppa, M. Hachey

Study supervision: J.H. Rowland

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.