Proceedings of a Workshop
Achieving Excellence in Cancer Diagnosis
Proceedings of a Workshop—in Brief
On October 6, 2021, the Board on Health Care Services of the National Academies of Sciences, Engineering, and Medicine hosted a virtual workshop to examine the current science and research opportunities for improving the diagnosis of cancer within the U.S. health care system.1 The workshop highlighted the patient experience, the diagnostic process, cancer epidemiology, approaches to mitigate disparities and promote equity in diagnosis, novel diagnostic strategies and tools, and strategies to improve the overall system of cancer diagnosis. This workshop was the third in a series on diagnostic excellence funded by the Gordon and Betty Moore Foundation.2 This Proceedings of a Workshop—in Brief highlights the presentations and discussions that occurred at the workshop.3
Daniel Yang of the Gordon and Betty Moore Foundation described the foundation’s framework of diagnostic excellence as embracing the six domains of health care quality in the Institute of Medicine report Crossing the Quality Chasm: A New Health System for the 21st Century: the diagnosis “must be safe, timely, effective, efficient, equitable, and patient centered” (IOM, 2001). He added that while all six domains are important, they are frequently in tension with one another and often require difficult trade-offs to achieve diagnostic excellence. For instance, while new diagnostic technologies may improve the accuracy of a cancer diagnosis, they are also likely to increase costs that may result in inequitable access, he said. Yang noted that cancers are one of the clinical areas4 that the Moore Foundation has selected as priorities because errors, delays, and inefficiencies in diagnosis in these areas can be especially harmful for patients and are responsible for a disproportionate share of serious patient harm and preventable death (Newman-Toker et al., 2019). He emphasized the importance of diagnostic excellence in cancer by noting that missed and delayed diagnoses in cancer are both common and consequential, and they are the number one cause of medical malpractice lawsuits in the United States. Andrew Bindman, chief medical officer and executive vice president of Kaiser Permanente, added that cancer differs from the other priority clinical areas (sepsis and acute cardiovascular events) because there are two major pathways to cancer diagnosis: screening of the asymptomatic population and symptom-based diagnosis. Ad-
1 The workshop agenda and presentations are available at https://www.nationalacademies.org/event/10-06-2021/achieving-excellence-in-cancer-diagnosis-a-workshop (accessed January 11, 2022).
2 More information about the workshop series is available at https://www.nationalacademies.org/our-work/advancing-diagnostic-excellence-a-workshop-series (accessed January 11, 2022).
3 This Proceedings of a Workshop—in Brief is not intended to provide a comprehensive summary of information shared during the workshop. The information summarized here reflects the knowledge and opinions of individual workshop participants and should not be seen as a consensus of the workshop participants, the planning committee, or the National Academies of Sciences, Engineering, and Medicine.
4 The other clinical areas include sepsis and cardiovascular events.
ditionally, the setting of diagnosis is commonly in primary care, as opposed to the emergency department, where it is less likely to be an acute event, and the timeline to diagnosis is often longer.
THE PATIENT EXPERIENCE
Tamika Felder is a 20-year survivor of cervical cancer, who founded Cervivor to empower individuals affected by cervical cancer through spreading awareness, education, and support. While she did have family support and a good medical team, she explained that there were things that fell through the cracks during her experience with cancer diagnosis, including lack of cultural understanding. Additionally, she noted that the wait time to receive her cancer diagnosis was stressful, and she emphasized that the timely communication of a cancer diagnosis is extremely important to patients. When she was diagnosed, she was scared and felt as if she had lost her voice within her care team; in order to reclaim her voice, Felder found that writing down questions and thoughts to provide to a clinician was helpful. She stressed the importance of patients being proactive and putting themselves first. She also highlighted the importance of establishing a support network of family, friends, and advocacy organizations, and of empowering patients to do their own research about diagnosis and treatment. In addition, Felder spoke about the power of patient stories in helping newly diagnosed patients and in “bringing the data to life” for clinicians and researchers. In reflecting on her own story, she described being diagnosed at the age of 25 and losing her fertility through a radical hysterectomy. At the time of her diagnosis, Felder’s insurance did not cover fertility preservation. Many years later, she was able to help change legislation in her state to ensure that future patients would not have that experience. “That is what being a patient advocate is all about…it is using your existence, your story to make a difference,” said Felder.
Gwen Darien, executive vice president at the National Patient Advocate Foundation, is a long-time patient advocate and three-time cancer survivor who shared her experiences with being diagnosed with cancer. Her first diagnosis was a delayed diagnosis of non-Hodgkin’s lymphoma approximately 25 years ago. As a young adult, seemingly healthy, she explained that her clinician dismissed her description of the severity of her symptoms and did not trust what she was saying. One of the first principles of diagnostic excellence is trust on both sides, she argued, and while she trusted her clinician, her clinician did not trust her. She added that delayed cancer diagnoses are intertwined with “access to equitable care, access to quality care, and access to ability to take time off,” as well as with patient fear of cancer diagnosis. After her first cancer diagnosis, she became a patient advocate to advance public dialogue in the hope that others would have a better experience than she did.
In the early days of advocacy, she said there were often gaps in care for people going through the screening and diagnostic process and being connected to treatment following diagnosis. For example, in 1990, the Centers for Disease Control and Prevention started a screening program for breast and cervical cancer, but it was not until 2000 that access to treatment was provided through Medicaid. Darien noted the challenges with cost and care coordination when discussing her breast cancer and endometrial cancer diagnoses. She described the high costs of screening and clinical care for endometrial cancer and her primary care provider’s lack of knowledge on the appropriate breast cancer screening considering her history with non-Hodgkin’s lymphoma treatment. To achieve excellence in cancer diagnosis, Darien said that there is a need to examine “the cost to patients, families, and communities when somebody is diagnosed,” and she cautioned that disparities in screening, diagnosis, and treatment “will persist until we deal with the cost of care in the fullest sense of the word, meaning not just clinical, but also cost of daily living, cost of returning to health and wellness.” She emphasized the need to ensure equitable, affordable health care for all, noting that while diagnostic excellence is important, it should be examined in a holistic way along with cancer treatment and care, including psychosocial needs.
OVERVIEW OF THE DIAGNOSTIC PATHWAYS FOR CANCER
Chyke Doubeni, professor at the Mayo Clinic, discussed the diagnostic pathway through screening, which aims to reduce cancer mortality through the early detection of cancer and precancer in asymptomatic populations. The World Health Organization (WHO) estimates that 30 to 50 percent of deaths from cancer could be prevented through early diagnosis and prompt and appropriate treatment, he said (WHO, 2022). However, many cancers are still being detected at a late stage when treatment is less effective. He also noted that effective screening tests are not available for most cancer types. Doubeni described the principles of screening, emphasizing that the tests should be safe, accurate, effective, acceptable to patients, and feasible to implement, with a clear diagnostic pathway when the screening test results are abnormal, and also accessible to all populations (Wilson et al., 1968). He emphasized the importance of ensuring that the potential benefits of a screening test outweigh the potential harms, noting that the U.S. Preventive Services Task Force uses an analytic framework to assess both the benefits (i.e., reducing morbidity and mortality) and harms (i.e., from the test itself, like perforation from colonoscopy; overdiagnosis; and false positives) and recommends
screening when net benefit is demonstrated. Cancer screening is a multistep process, he said, with potential harms at each stage, which is why timely and continuous assessment is crucial to improving the quality of cancer screening. A 2019 study found that individuals with abnormal colorectal cancer screening results who did not receive follow-up testing had a seven-fold higher risk for colorectal cancer death (Doubeni et al., 2019). He listed some best practices in screening delivery, including
- goal to achieve high screening rates across populations using Healthcare Effectiveness Data and Information Set (HEDIS) measures;
- timely initiation of screening, and rescreening at appropriate intervals;
- high-quality screening and timely follow-up when results are abnormal; and
- timely delivery of treatment when cancer is detected.
In conclusion, he suggested using a health equity framework to improve the cancer screening process, considering community integration as a foundational element, examining social determinants of health to address disparities in cancer screening and diagnosis, and understanding that screening is a multistep process.
Elizabeth Sarma, program director at the National Cancer Institute, reviewed the diagnostic pathway of symptom-detected cancers in the United States, reiterating that the ultimate goal is to detect all cancers at an early stage when it is more amenable to treatment. Even under ideal conditions, only one-third of cancers are potentially detectable through screening, leaving two-thirds to be detected through other pathways (Sarma et al., 2020).
While the United States lacks the population data used in other countries to obtain the proportion of cancers diagnosed in different pathways, some of the most compelling evidence comes from the International Cancer Benchmarking Partnership (ICBP), which explores factors that can influence differences in cancer survival across eight countries with similar health care systems. Across jurisdictions in the ICBP, it found 82 percent of colorectal cancer diagnoses occurred after symptomatic presentation and just 16 percent were detected via screening (Weller et al., 2018). So even in countries that have dedicated screening programs, these cancers are not commonly detected early by screening, she said. Instead, research suggests that most cancers are detected and diagnosed after individuals present with symptoms; thus, another strategy for early diagnosis is to identify symptomatic disease at the earliest possible opportunity and link it to a diagnosis and treatment without delay (WHO, 2017).
Sarma mentioned evidence exists that reduced time to diagnosis of symptomatic cancer is associated with better clinical outcomes, better patient-reported outcomes and care satisfaction, lower treatment cost, and improved care quality (Neal et al., 2015; Newman-Toker et al., 2019). However, she says challenges in this pathway include appraisal of signs and symptoms, and the commonalities among symptoms of cancer and other conditions, as well as a lack of tools for diagnostic evaluation and effective communication and coordination across multiple medical disciplines to get a timely diagnosis. Sarma noted that most of the evidence on early diagnosis from symptoms is coming from other countries such as the United Kingdom (UK), and more information could be elucidated in the United States by collecting data to describe prediagnosis care and diagnostic pathways in U.S. health care systems. Further work is needed to understand symptom epidemiology and which symptoms, signs, and test results are helpful in identifying patients at higher risk for specific cancers, Sarma concluded.
Fiona Walter, professor at Queen Mary University of London, shared lessons from the UK experience with cancer detection and diagnosis in primary care. She described a wake-up call in 2000 when a number of studies reported that the UK lagged behind neighboring countries with similar health care systems in survival outcomes for breast, colorectal, lung, and ovarian cancers. She reviewed the UK’s public health care system, commonly referred to as the National Health Service, which is available to all residents and paid for by taxation. Multiple policy approaches were implemented to improve cancer diagnosis, starting with the establishment of the 2-week referral standard in which patients are offered appointments with relevant specialists within 2 weeks of being referred by a primary care clinician. Walter briefly touched upon the three national screening programs focused on bowel, breast, and cervical cancer, reporting that English data from a 2017 National Cancer Diagnosis Audit showed that only 7 percent of cancer diagnoses came from the screening pathway, while 64 percent came through primary care after patients presented with symptoms (Swann et al., 2018).
For the past decade, Walter said, the UK has focused on identifying clinical features associated with specific cancers to improve diagnosis. National Health Service primary care data, which is a part of the General Practice Research Database and Clinical Practice Research Datalink, were used to develop lists of symptoms and to determine positive predictive values of the lists for identifying patients with cancer.5 These data were used to inform the 2015
5 Positive predictive value refers to the probability that following presentation of a symptom, an individual will be diagnosed with cancer.
UK clinical guidance around recognition and referral for suspected cancers (NICE, 2015). Walter also shared another approach to developing cancer risk prediction tools, using primary care data and algorithms that predict individual level of cancer risk based on multiple factors and symptoms. Embedded in 50 percent of the general practice system nationally as interactive calculators, Walter explained that these tools allow patients with symptoms and their clinicians to understand risk of cancer prior to diagnosis.
She also shared an example of the use of a blood test for CA125 as a marker for ovarian cancer, saying that a recent evaluation of its performance in a large dataset from English general practice found it to be more predictive of cancer in symptomatic women than previously thought (Funston et al., 2020). Walter notes predictive models have since been developed to estimate probability of cancer in symptomatic patients and are now being used in routine UK primary care. Finally, she described a diagram demonstrating the potential for improvements across the diagnostic pathway in the UK. In particular, Walter stressed that “we have a great potential for better cancer outcomes” by detecting more cancers through screening programs, encouraging earlier help-seeking for symptoms, and ensuring earlier referral for diagnostic tests.
The panel discussion centered around the challenges posed by the multistep process of cancer screening and diagnosis, as well as the inefficiency of using tests that screen for a single type of cancer and the potential future of bundling screenings for multiple cancers with a single test. Doubeni and Walter agreed that bundling screenings could be beneficial and that some early methods are showing promise, but they stressed that there is still a large amount of work to be done to achieve this goal, given the current practices, and to ensure such a test is effective in reducing cancer mortality and is acceptable to patients. Doubeni added that it is important to be mindful of undue harms caused by screening. Sarma emphasized that one size will not fit all in the development and execution of diagnostic pathways, so it is important to understand the population needs and the cancer care context before launching an intervention. Walter added that easier access to diagnostics can improve patient triage more effectively within primary care.
CANCER EPIDEMIOLOGY AND STRATEGIES TO MITIGATE DISPARITIES AND PROMOTE EQUITY
Caroline Thompson, associate professor at the University of North Carolina, Chapel Hill, focused her remarks on the epidemiology of cancer and the impact of disparities on cancer outcomes for underserved populations, such as low-income, rural, or underinsured populations. Using lung cancer as an example, she demonstrated how diagnostic disparities affect cancer outcomes across populations. Lung cancer prognosis can be favorable when a patient receives an early-stage diagnosis and timely treatment, but Thompson said that considerable disparities exist among populations because the stage of disease at diagnosis is not evenly distributed across populations. For example, in 2016, nearly 70 percent of White patients and only 57 percent of Black patients were diagnosed at an early stage (Offor et al., 2020).
Access to treatment is similarly plagued with disparities, she said, resulting in poorer outcomes for Black patients with lung cancer compared to White patients with lung cancer. She explained that epidemiologists are still trying to identify the sources of disparities in detection, diagnosis, and treatment, which are reflected in access to care and the quality of care. Some of these sources include participation in screening, timely diagnosis, and timely and appropriate treatment. Thompson also emphasized the importance of identifying the drivers of such disparities in order to guide targeted interventions to reduce inequities. These drivers include geography (e.g., whether a person lives in a rural or urban area), cultural and social factors, and economic factors. She described barriers to ensuring early diagnosis and prompt treatment, including lack of patient awareness and low health literacy, stigma, and limited access to care, as well as financial barriers and transportation barriers. Thompson concluded: “Within the health care system, accurate diagnosis demands prompt attention to patient complaints, accurate differential diagnosis and, importantly, continuity of care.”
Kathy Tossas, assistant professor at the Virginia Commonwealth University, described strategies to mitigate racial, ethnic, and geographic disparities in cancer. She focused on the Cancer Disparities Progress Report of the American Association for Cancer Research (Sengupta and Honey, 2020) and highlighted some of the key strategies to mitigate disparities, such as enhancing diversity, equity, and inclusion in the cancer workforce; increasing federal funding to study and address disparities; collaborative and coordinated resourcing such as public–private partnerships; and public policy to support a healthy diet, smoking cessation, cancer screening access and quality, and access to clinical trials. Tossas stressed that while a disadvantage might be inherent to an individual, it is important to view the overarching structural barriers and infrastructure as the problems that need to be addressed. She proposed the following changes:
- Collect integrated, decentralized, and democratized data and use it to drive equity.
- Acknowledge and redistribute power through allyship and sponsorship.
- Shift the focus from trust to trustworthiness through inward reflection.
“We need to identify and understand that we come to the table with different experiences…before we begin to create our intervention,” Tossas said.
Blase Polite, professor at the University of Chicago, highlighted system-level strategies to reduce disparities and promote equity, emphasizing the importance of public health and policy in achieving health equity. For example, in states that did not expand their Medicaid program under the Affordable Care Act of 2010, the majority of individuals in the coverage gap were people of color. This is important in regard to improving health disparities, he said, because Medicaid expansion results in increased screening and increased early-stage cancer diagnosis with better prognoses. In addition, studies have found that patients with better insurance coverage are diagnosed at an earlier stage of cancer and have better access to cancer surgery (Han et al., 2018).
Polite also described the C5 initiative, developed in 2003 by the New York City Department of Health and Mental Hygiene to increase awareness of and access to colonoscopy. The initiative targeted health messages about the benefits of screening to underserved areas, coupled them with open access colonoscopies,6 and provided patient navigation services. By 2013, he said, they increased colonoscopy rates from 40 percent of the eligible population to more than 70 percent, and essentially eliminated the disparities between the Black and White populations in New York City (Itzkowitz et al., 2016). Polite also described the Delaware Cancer Consortium Plan, which prioritized reduction of disparities through a comprehensive statewide colorectal cancer screening program.7 In 2001, only 15 percent of Black patients were diagnosed with early-stage colorectal cancer, but by 2009, 50 percent were diagnosed at an early stage (Grubbs et al., 2013). Most importantly, implementation of the plan led to decreased cancer incidence rates, which indicates increased detection and removal of precancerous polyps, and decreased cancer mortality rates for Black patients, achieving parity with other populations. Polite concluded by highlighting the following strategies to mitigate disparities: facilitating early diagnosis through access to screening, providing care navigation services for high-risk patients, and holding state Medicaid programs and Medicaid managed care organizations accountable for health equity outcomes.
In a panel discussion, speakers commented on the opportunities to proactively ensure that implementation of new interventions does not perpetuate structural inequities or discriminate against underserved populations. Tossas called for a focus on diversity throughout the care continuum and for revising established policies to make them more equitable. Polite expressed concern about a lack of patient diversity in clinical trials, which could widen disparities in treatment outcomes. Thompson added that in order to understand the patterns of how patients with cancer navigate the health care system, more data linkages are needed between Medicaid and cancer registries like the Surveillance, Epidemiology, and End Results (SEER) Program.
NOVEL STRATEGIES AND TOOLS FOR CANCER DIAGNOSIS
James Ford, professor at Stanford University School of Medicine, gave an overview of precision medicine in cancer, and how molecular tumor profiling and targeted therapeutics are improving outcomes. By identifying the detailed molecular changes in an individual tumor, clinicians can select therapies most likely to be effective for each patient. Ford also described the development of a “molecular tumor board,” a multidisciplinary group of experts who work together to integrate the complex information derived from tumor profiling to help identify the appropriate drug for a patient. He described several opportunities for using individualized tumor molecular profiles to improve diagnosis and treatment and to understand drug resistance mechanisms (Hainsworth et al., 2018). But challenges still remain, he noted, such as identifying the right profiling strategies and assessing their effect on patient outcomes, the difficulties in conducting clinical trials because of tumor heterogeneity, the costs of testing and its effect on patient access, and how to develop and implement standards of care for testing and treatment. Ford noted that next-generation DNA sequencing is being used both for genetic analysis of germline DNA to identify hereditary cancer syndromes as well as genomic analysis of somatic DNA from tumors. Future approaches may include doing combined genetic and genomic analysis of germline and somatic DNA, incorporating RNA sequencing, and assessing circulating tumor DNA.
Victor Velculescu, professor at Johns Hopkins University School of Medicine, discussed cell-free tumor DNA fragments and their potential use for early detection of cancer. Globally, about 14 million people are diagnosed with cancer and 8 million people die of cancer each year, he reported, and those numbers are projected to rise to 25 million new cases and 15 million annual deaths by 2035 (WHO, 2021). He reiterated points made by previous speakers, saying that early diagnosis can lead to much more favorable outcomes, and suggesting that if effective early detection approaches were fully implemented, they could potentially save four to six million lives each year. While tests exist for early detection, the access to and use of these tests is suboptimal, so he discussed the possibility of developing more
6 Open access colonoscopies do not require the patient to see a gastrointestinal doctor.
7 For further information on the Delaware Cancer Consortium, see https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3661932 (accessed February 1, 2022).
cost-effective and high-performing methods for early detection. His work has focused on an approach called the DNA Evaluation of Fragments for Early Interception (DELFI), based on the finding that individuals with cancer release abnormal DNA fragments into the blood.
Research indicates that healthy individuals have very similar cell-free tumor DNA fragmentation profiles, but individuals with cancer have widely disparate profiles. Velculescu and his colleagues have applied the DELFI approach and machine learning to this data to develop a potential new test for early detection of seven different cancer types, he said (Cristiano et al., 2019). Because lung cancer is so deadly and current screening approaches are poorly utilized, Velculescu said they have focused on using a DELFI approach for detection of lung cancer across different stages and subtypes (Mathios et al., 2021),8 raising the possibility of a widely accessible and cost-effective test as a prescreen to the current screening modality of lung cancer. Velculescu also noted that fragmentation profiles are different among cancer types and that machine learning can be used to identify a cancer’s tissue of origin (Cristiano et al., 2019).
Maryellen Giger, professor at the University of Chicago, described novel technologies and strategies that aim to optimize cancer diagnosis using imaging, cautioning that the benefit of medical imaging depends on both the quality of the imaging and the quality of the interpretation. The goals for artificial intelligence (AI) in oncologic imaging, she explained, are to develop computer-aided technologies for detecting and diagnosing cancer early and to improve the efficiency and workflow of medical imaging interpretation. She said AI can be applied to many different cancer imaging tasks, including risk assessment, screening, diagnosis, and treatment planning. Multiple studies, though limited in scope, have shown that the performance of radiologists in characterizing and diagnosing lesions improves when using the computer aid as a second reader (Rodriguez-Ruiz et al., 2019; Shimauchi et al., 2011). She mentioned the combination of human-engineered AI algorithms and deep transfer learning led to statistically improved performance of the AI algorithm.
One important gap in the development of AI is the lack of data on diverse patients that are representative of the population receiving care. When describing the future of AI, Giger said integrated, multimodality, multitask AI offers a potential means to improve access to health care and reduce health disparities. For instance, she suggested that AI can be used when access to expert radiologists is limited. She also described the need for multiomics data sets for both discovery and development of clinical biomarkers that can be used to assess tumor susceptibility and response to treatments over time. She also emphasized the importance of ethical and unbiased data collection methods to develop trustworthy AI algorithms.
Andrew Beck, chief executive officer of PathAI, described similar novel technologies and strategies to optimize diagnosis in pathology. Pathologic analysis answers questions such as whether the patient has cancer, how advanced the disease is, and which treatments are likely to be effective. He described the experience of a cancer epidemiologist who was diagnosed with cancer herself and sent tissue samples to multiple pathologists who all gave differing responses. Seeing this level of diagnostic variability led her to conduct a study focusing on observer accuracy and reproducibility. Overall, researchers found concerningly low levels of agreement for many different types of skin biopsies, with one biopsy resulting in 18 different diagnoses from 36 different pathologists (Elmore, et al., 2017). This diagnostic variability has been a core problem in pathology for decades, Beck said, but until very recently, technology systems have not worked well enough to improve clinical practice. He said deep learning for computer image analysis has shown remarkable progress in recent years, with the best systems now outperforming humans. This is increasingly important in cancer diagnostics because the landscape of precision medicine is becoming more and more complex, and there is a need for highly accurate, reproducible, and data-driven diagnostics to precisely match patients with a therapy that is most effective for their disease. As an example, he discussed the challenges with traditional pathology review of PD-L1 immunohistochemistry, which involves manually estimating the amount of PD-L1 protein expressed on cancer cells (Rimm at al., 2017). Beck added that AI-driven pathology has the potential to transform the field by enabling more accurate diagnosis in clinical trials and drug development.
With more than 100 algorithms for diagnosis cleared by the FDA, the speaker panel was asked for suggestions on how to translate research into clinical practice. Giger said it is important to first define the clinical aim or target population, and then use that to inform AI development and study design. She added that more data is needed to train and validate algorithms, and she suggested that creating open data repositories for testing and validating algorithms could expedite development and regulatory review and get effective technologies to the public faster. Velculescu highlighted the importance of collaboration among experts in different fields, such as biostatistics, epidemiology, clinical oncology, and pathology, to accelerate progress. Ford added that increased sharing of genomic data across institutions is important for these collaborations.
8 In the spring of 2021, the company he founded launched its first prospective, case-control study to measure the DELFI test’s performance in a group of patients already known to have lung cancer and a matched cohort of healthy patients.
IMPROVING THE SYSTEM FOR CANCER DIAGNOSIS
Sallie Weaver, program director at the National Cancer Institute, discussed strategies for improving cancer care coordination in the diagnostic process, which often involves several specialties and multiple care settings. Historically, the coordination requirements for diagnosis were not emphasized in workflows, she explained. For example, showing a diagram of the “idealized pathway for breast cancer screening” from the early 2000s, she noted its simplicity as well as some of the gaps in follow-up care. The diagnostic pathway map in 2021 is much more complex, but she reiterated that efforts are still needed to improve coordination among multidisciplinary care teams, and said that coordination strategies should evolve as new technologies become integrated into standard clinical practice. She said very few studies have focused on coordination during diagnosis, and she emphasized the importance of employing a system lens to see the whole process, including interdependencies and factors that influence the process at multiple levels (the patient, clinician, care team, care setting, and health system).
One limitation to current tools is they are not individualized and do not reflect care tasks that vary by patient preferences or insurance status, Weaver said, but there are ways to address this. She described the 4R intervention that uses personalized care sequences to visualize key care tasks including the timing, sequencing, and roles of care team members. Another challenge is that coordination is being defined in different ways, and outcomes are being measured differently. She explained that this makes it difficult to compare results across studies and contributes to variable or contradictory estimates of intervention effectiveness. One path forward is to clearly differentiate coordination mechanisms (tools and strategies used) from coordination behaviors (teamwork skills and other behaviors used in communicating) and to focus on shared accountability.
Electra Paskett, professor at the Ohio State University, shared strategies for improving cancer screening and diagnosis developed through her work in rural Appalachia. She described the region as having nearly 25 million residents across 13 states, with both urban and rural areas, but less racial diversity, higher rates of poverty, and lower education rates compared to the rest of the country. Cancer is the leading cause of death in Appalachia, with many factors contributing to the region’s health disparities such as lack of medical care facilities and health care providers, lower socioeconomic status, few specialists, and travel required for advanced care. People in the region are often reluctant to talk about cancer, she noted, making it difficult to increase awareness and access to care. Paskett shared several vignettes illustrating system-level problems, such as a lack of capacity for follow-up care when women receive an abnormal result from a cervical pap smear, a dearth of mammography facilities across seven counties, and a lack of capacity in colorectal cancer screening and follow-up. Paskett described how a community cancer coalition obtained a grant to establish a free clinic where women had access to exams and treatment to improve follow-up care after an abnormal cervical cancer screening.
In response to the dearth of mammography facilities, community partners also facilitated access to a mobile mammography van, and a continuum of care navigation program was instituted in which local community health workers go into the community to educate women about the need for screening. A navigator works with women who test positive to ensure follow-up through diagnostic resolution and treatment. A multilevel focus on community, clinics, and care providers was used to improve colorectal screening as well. Specifically, implementation science methods were used to determine the evidence-based intervention to implement at each level in the community health center, with the goal of disseminating successful strategies to the entire health system. She concluded with the take-home message that local solutions solve local problems. Despite barriers, clinicians want to deliver the best care, and the community can be receptive to messages about health if they are shared in the right manner. To reach communities, Paskett emphasized the “need to meet them where they are,” develop local capacity, and leverage patient navigation and implementation science.
Jasmin Tiro, professor at the University of Texas Southwestern Medical Center, described best practices in cancer screening and early detection across cancer types and health care systems. Because of new technologies, there have been rapid changes in diagnostic evaluation after screening. She shared a conceptual model for a screening process, displaying where screening results warrant immediate evaluation or ongoing surveillance. She found that in a PROSPR cervical cohort of more than 800,000 women,9 65 percent had no information on cervical cancer risk and screening history. When these women have an abnormal test result, she explained, it is much more difficult to apply management guidelines and determine next steps in the process. Tiro said they identified two failures in the diagnostic care pathway through their PROSPR consortium: difficulty tracking referrals, either because appointments were cancelled by the system or the patient, and patients not understanding the reason for the referral and thus not
9 Population-based Research to Optimize the Screening Process (PROSPR) is a network that is conducting research to better understand how to improve the cancer screening process (recruitment, screening, diagnosis, referral for treatment) in community health care settings. For more, see https://healthcaredelivery.cancer.gov/prospr (accessed January 12, 2022).
prioritizing that appointment in the face of competing life demands, highlighting the importance of patient-centered, closed-loop communication. But when care teams do not have shared clinical information systems and results in structured fields, it is difficult to optimize workflows and implement automated health information technology (IT) solutions.
Tiro noted a best practice is to calculate timeliness quality metrics for diagnosis of each cancer type, highlighting that currently only the National Breast and Cervical Cancer Early Detection Program requires reporting of quality metrics on diagnostic evaluation. To highlight other challenges in the diagnostic pathway, Tiro shared results from a clinician survey that showed only 69 percent identified the correct management plan for a patient vignette, with only 42 percent confident in their decision. Additionally, if scheduling systems are not open one year in advance, then staff will have to recontact and schedule patients for annual follow-up, leading to more opportunity for gaps in care. In summary, she stated that many safety-net health care settings have limited resources and competing demands, leading to an inconsistent availability of cancer screening programs supported by public payers. Safety-net health care settings also have limited ability to invest in staff to report information to registries for population health improvement. “We cannot optimize diagnostic care delivery and produce equitable health outcomes with this inconsistent coverage,” Tiro emphasized.
During the panel discussion, when asked about features of systems that enable these interventions to work well, Tiro responded that stakeholder buy-in and motivations were key. Paskett added that patient navigation, which currently is not reimbursable, would also be a valuable addition in every setting. She also noted that something as simple as a staff tutorial on the electronic health record (EHR) capabilities at the system level could allow clinicians to more effectively manage follow-up care. Weaver stated that patient navigators are often using IT platforms that are not well integrated into the overall EHR, so time is spent trying to connect pieces of data instead of navigating patients through the care continuum. Implementing new management guidelines for different cancer screenings also take time and effort, Paskett said. Finally, Sarah Kobrin from the National Cancer Institute, highlighted the discrepancy in patient cost-sharing for screening and diagnostic tests—noting that if an individual comes in for screening without symptoms, then insurance coverage without patient cost-sharing is mandated through the Affordable Care Act, but diagnostic testing for a patient with symptoms is not included in this policy and thus can be very costly for patients. Tiro agreed that better coverage of diagnostic tests is a critical need in the country. Finally, Weaver added that staffing and resource issues are often a challenge for rural locations, so bringing services directly to patients or using care models to connect patients to specialists are other ways to address this gap.
VISION FOR THE FUTURE
In a final session, workshop panelists reflected on their visions for diagnostic excellence in cancer diagnosis. Mia Levy, chief medical officer at Foundation Medicine, emphasized the time to first treatment as a quantitative outcome measure that many cancer centers are beginning to use. But there are many challenges that can lead to delays in the diagnostic pathway, as well as delays between diagnosis and treatment initiation, with requirements for insurance authorization being a key factor. Levy also discussed the importance of incorporating the delivery of screening care into a learning health care system model, in which data collection is a standard component of care and data are used to inform clinical decision making and to drive improvements in the quality of care.
Tamika Felder from Cervivor stressed the importance of empowering patients to understand and listen to what their bodies might be telling them—even if in a whisper—in order to seek timely care. This also includes building trust in a system that has historically not served all populations in an equitable way. Focusing on the poor outcomes for patients with delayed or missed diagnoses, Luke Sato, senior vice president and chief medical officer of CRICO, the malpractice insurer of the Harvard Medical Institutions, said that cancer misdiagnosis cases make up a majority of medical malpractice cases. To combat this, CRICO is working with all hospitals they insure to develop safety strategies, with the goal of improving care coordination and EHR interoperability to close the “referral loop” throughout their network. Gordon Schiff, associate director of Brigham and Women’s Hospital, added that delivering timely care at every stage is essential, especially for patients who are symptomatic. He suggested the health system also needs to hear more from patients about their experience and learn from what they have been through.
More suggestions from workshop speakers to improve excellence in cancer diagnosis are outlined in Box 1.
SUGGESTIONS FROM INDIVIDUAL WORKSHOP PARTICIPANTS TO ACHIEVE EXCELLENCE IN CANCER DIAGNOSIS
Improving Care Delivery Innovations
- Develop a system that communicates to clinicians in real time when patients are lost to follow-up, and have a method to intervene (Polite, Sato).
- Incorporate screening care into a learning health care system in which routine data collection can improve decision support and contribute to data-driven national guidelines (Levy).
- Develop community diagnostic hubs with easy access to diagnostic testing (Walter).
- Measure time to first treatment to improve cancer diagnosis and hold payers accountable for their role in authorization delays (Levy, Polite).
Policy Opportunities and Operational Changes
- Expand Medicaid coverage in areas with known health disparities to optimize early diagnosis and produce more equitable health outcomes (Polite, Tiro).
- Synchronize insurance coverage with changes in clinical guidelines and reduce patient cost-sharing for diagnostic testing (Kobrin, Paskett, Tiro).
- Make better use of Medicaid data through linkages to cancer registries (Thompson).
Emphasizing Patient-Centered Care
- Incorporate the voices of patients, caregivers, patient advocates, and patient navigators throughout the cancer care continuum (Darien, Felder).
- Improve timely communication of a cancer diagnosis to patients (Darien, Felder).
- Examine relationships among access to care, timeliness of care, and delayed diagnoses (Darien, Doubeni).
- Systematically gather data on patients with symptoms that may indicate cancer (Schiff).
- Develop systems to retrospectively learn from patients who have experienced errors and delayed diagnosis (Schiff).
Focusing on Equity
- Increase equitable access to screening and diagnosis (Felder).
- Improve public health messaging and communications for cancer screening (Darien).
- Use a health equity framework to improve the cancer screening process by incorporating community perspectives and social determinants of health (Doubeni).
- Incorporate diverse community voices and leverage patient navigation and implementation science to improve cancer screening and diagnosis (Felder, Paskett, Tossas).
- Collect more granular data to identify and address disparities (Polite, Thompson, Tossas).
- Hold state Medicaid programs and Medicaid managed care organizations accountable for health equity outcomes (Polite).
Opportunities for Future Research Initiatives
- Collect population data on the epidemiology of cancer diagnosis to describe prediagnosis care and diagnostic pathways in U.S. health care systems (Sarma, Thompson).
- Evaluate the potential benefits and harms of novel multicancer screening tests (Doubeni, Walter).
- Invest in shared multiomics data sets for both discovery and development of clinical biomarkers (Giger).
BOX 1 CONTINUED
- Continue testing multilevel interventions, and develop novel measures for diagnostic care coordination, for both asymptomatic and symptomatic patients (Weaver).
NOTE: This list is the rapporteurs’ summary of points made by the individual speakers identified, and the statements have not been endorsed or verified by the National Academies of Sciences, Engineering, and Medicine. They are not intended to reflect a consensus among workshop participants.
Cristiano, S., A. Leal, J. Phallen, J. Fiksel, V. Adleff, D. C. Bruhm, S. Ø. Jensen, J. E. Medina, C. Hruban, J. R. White, D. N. Palsgrove, N. Niknafs, V. Anagnostou, P. Forde, J. Naidoo, K. Marrone, J. Brahmer, B. D. Woodward, H. Husain, K. L. van Rooijen, M. W. Ørntoft, A. H. Madsen, C. J. H. van de Velde, M. Verheij, A. Cats, C. J. A. Punt, G. R. Vink, N. C. T. van Grieken, M. Koopman, R. J. A. Fijneman, J. S. Johansen, H. J. Nielsen, G. A. Meijer, C. L. Andersen, R. B. Scharpf, and V. E. Velculescu. 2019. Genome-wide cell-free DNA fragmentation in patients with cancer. Nature 570(7761):385-389.
Doubeni, C. A., S. A. Fedewa, T. R. Levin, C. D. Jensen, C. Saia, A. M. Zebrowski, V. P. Quinn, K. A. Rendle, A. G. Zauber, T. A. Becerra-Culqui, S. J. Mehta, R. H. Fletcher, J. Schottinger, and D. A. Corley. 2019. Modifiable failures in the colorectal cancer screening process and their association with risk of death. Gastroenterology 156(1):63-74.e6.
Elmore, J. G., R. L. Barnhill, D. E. Elder, G. M. Longton, M. S. Pepe, L. M. Reisch, P. A. Carney, L. J. Titus, H. D. Nelson, T. Onega, A. Tosteson, M. A. Weinstock, S. R. Knezevich, and M. W. Piepkorn. 2017. Pathologists’ diagnosis of invasive melanoma and melanocytic proliferations: Observer accuracy and reproducibility study. BMJ (Clinical research ed.) 357, j2813.
Funston, G., W. Hamilton, G. Abel, E. J. Crosbie, B. Rous and F. M. Walter. 2020. The diagnostic performance of CA125 for the detection of ovarian and non-ovarian cancer in primary care: A population-based cohort study. PLOS Medicine 17(10).
Grubbs, S. S., B. N. Polite, J. Carney Jr., W. Bowser, J. Rogers, N. Katurakes, P. Hess, and E. D. Paskett. 2013. Eliminating racial disparities in colorectal cancer in the real world: It took a village. Journal of Clinical Oncology 31(16):1928-1930.
Hainsworth, J. D., F. Meric-Bernstam, C. Swanton, H. Hurwitz, D. R. Spigel, C. Sweeney, H. Burris, R. Bose, B. Yoo, A. Stein, M. Beatty, and R. Kurzrock. 2018. Targeted therapy for advanced solid tumors on the basis of molecular profiles: Results from MyPathway, an open-label, phase IIa multiple basket study. Journal of Clinical Oncology 36(6):536-544.
Han, X., K. R. Yabroff, E. Ward, O. W. Brawley, and A. Jemal. 2018. Comparison of insurance status and diagnosis stage among patients with newly diagnosed cancer before vs after implementation of the Patient Protection and Affordable Care Act. JAMA Oncology 4(12):1713-1720.
IOM (Institute of Medicine). 2001. Crossing the quality chasm: A new health system for the 21st century. Washington, DC: National Academy Press. https://doi.org/10.17226/10027 (accessed February 12, 2022).
Itzkowitz, S. H., S. J. Winawer, M. Krauskopf, M. Carlesimo, F. H. Schnoll-Sussman, K. Huang, T. K. Weber, and L. Jandorf. 2016. New York citywide colon cancer control coalition: A public health effort to increase colon cancer screening and address health disparities. Cancer 122(2):269-277. https://doi.org/10.1002/cncr.29595 (accessed February 12, 2022).
Mathios, D., J. S. Johansen, S. Cristiano, J. E. Medina, J. Phallen, K. R. Larsen, D. C. Bruhm, N. Niknafs, L. Ferreira, V. Adleff, J. Y. Chiao, A. Leal, M. Noe, J. R. White, A. S. Arun, C. Hruban, A. V. Annapragada, S. Ø. Jensen, M.-B. W. Ørntoft, A. H. Madsen, B. Carvalho, M. de Wit, J. Carey, N. C. Dracopoli, T. Maddala, K. C. Fang, A.-R. Hartman, P. M. Forde, V. Anagnostou, J. R. Brahmer, R. J. A. Fijneman, H. J. Nielsen, G. A. Meijer, C. L. Andersen, A. Mellemgaard, . E. Bojesen, R. B. Scharpf, and V. E. Velculescu. 2021. Detection and characterization of lung cancer using cell-free DNA fragmentomes. Nature Communications 12(5060).
Neal, R. D., P. Tharmanathan, B. France, N. U. Din, S. Cotton, J. Fallon-Ferguson, W. Hamilton, A. Hendry, M. Hendry, R. Lewis, U. Macleod, E. D. Mitchell, M. Pickett, T. Rai, K. Shaw, N. Stuart, M. L. Tørring, C. Wilkinson, B. Williams, and N. Williams, and J. Emery. 2015. Is increased time to diagnosis and treatment in symptomatic cancer associated with poorer outcomes? Systematic review. British Journal of Cancer 112(1):S92-S107.
Newman-Toker, D. E., A. C. Schaffer, C. W. Yu-Moe, N. Nassery, A. S. S. Tehrani, G. D. Clemens, Z. Wang, Y. Zhu, M. Fanai, and D. Siegal. 2019. Serious misdiagnosis-related harms in malpractice claims: The “big three”—vascular events, infections, and cancers. Diagnosis 6(3):227-240.
NICE (National Institute for Health and Care Excellence). 2015. Suspected cancer: Recognition and referral. NICE guideline. https://www.nice.org.uk/guidance/ng12/resources/suspected-cancer-recognition-and-referral-pdf-1837268071621 (accessed October 25, 2021).
Offor, P., C. Obrochta, B. Schumacher, and C. Thompson. 2020. Trends in stage at diagnosis for lung cancer in the U.S., 2009-2016 [abstract]. Cancer Epidemiology, Biomarkers & Prevention 29(12 Suppl): abstract nr PO-199.
Rimm, D. L., G. Han, J. M. Taube, S. Y. Eunhee, J. A. Bridge, D. B. Flieder, R. Homer, W. W. West, H. Wu, A. C. Roden, J. Fujimoto, H. Yu, R. Anders, A. Kowalewski, C. Rivard, J. Rehman, C. Batenchuk, V. Burns, F. R. Hirschand, and I. I. Wistuba. 2017. A prospective, multiinstitutional, pathologist-based assessment of 4 immunohistochemistry assays for PD-L1 expression in non–small cell lung cancer. JAMA Oncology 3(8):1051-1058.
Rodríguez-Ruiz, A., E. Krupinski, J. J. Mordang, K. Schilling, S. H. Heywang-Köbrunner, I. Sechopoulos, and R. M. Mann. 2019. Detection of breast cancer with mammography: Effect of an artificial intelligence support system. Radiology 290(2):305-314.
Sarma, E. A., S. C. Kobrin, M. J. Thompson. 2020. A proposal to improve the early diagnosis of symptomatic cancers in the United States. Cancer Prevention Research 13(9):715-720.
Sengupta, R., and K. Honey. 2020. AACR Cancer Disparities Progress Report 2020: Achieving the bold vision of health equity for racial and ethnic minorities and other underserved populations. Cancer Epidemiology, Biomarkers & Prevention 29(10):1843.
Shimauchi, A., M. L. Giger, N. Bhooshan, L. Lan, L. L. Pesce, J. K. Lee, H. Abe, and G. M. Newstead. 2011. Evaluation of clinical breast MR imaging performed with prototype computer-aided diagnosis breast MR imaging workstation: Reader study. Radiology 258(3):696-704.
Swann, R., S. McPhail, J. Witt, B. Shand, G. A. Abel, S. Hiom, J. Rashbass, G. Lyratzopoulos, G. Rubin, and the National Cancer Diagnosis Audit Steering Group. 2018. Diagnosing cancer in primary care: Results from the National Cancer Diagnosis Audit. British Journal of General Practice 68(666):e63-e72.
Weller, D., U. Menon, A. Z. Falborg, H. Jensen, A. Barisic, A. K. Knudsen, R. J. Bergin, D. Brewster, V. Cairnduff, A. T. Gavin, E. Grunfeld, E. Harland, M. Lambe, R.-J. Law, Y. Lin, M. Malmberg, D. Tober, R. Neal, V. White, S. Harrison, and I. Reguilon. 2018. Diagnostic routes and time intervals for patients with colorectal cancer in 10 international jurisdictions; Findings from a cross-sectional study from the International Cancer Benchmarking Partnership (ICBP). BMJ Open 8(11).
WHO (World Health Organization). 2017. Guide to cancer early diagnosis.https://apps.who.int/iris/handle/10665/254500 (accessed February 12, 2022).
WHO. 2021. Cancer over time.https://gco.iarc.fr (accessed December 13, 2021).
WHO. 2022. Cancer. https://www.who.int/news-room/fact-sheets/detail/cancer (accessed January 11, 2022).
Wilson, J. M. G., G. Jungner, and the World Health Organization. 1968. Principles and practice of screening for disease. World Health Organization. https://apps.who.int/iris/handle/10665/37650 (accessed February 12, 2022).
DISCLAIMER: This Proceedings of a Workshop—in Brief was prepared by Jennifer Lalitha Flaubert, Ruth Cooper, Megan Snair, and Sharyl Nass as a factual summary of what occurred at the meeting. The statements made are those of the rapporteurs or individual workshop participants and do not necessarily represent the views of all workshop participants, the planning committee, or the National Academies of Sciences, Engineering, and Medicine.
*The National Academies of Sciences, Engineering, and Medicine’s planning committees are solely responsible for organizing the workshop, identifying topics, and choosing speakers. The responsibility for the published Proceedings of a Workshop—in Brief rests with the workshop rapporteurs and the institution. The planning committee comprises Andrew Bindman (Chair), Kaiser Permanente; Pascale Carayon, University of Wisconsin-Madison; Robert Carlson, National Comprehensive Cancer Network; Sarah Kobrin, National Cancer Institute; Alex Krist, VCU School of Medicine; Curtis Langlotz, Stanford University; Kathryn McDonald, Johns Hopkins University; Ritu Nayar, Northwestern University; Lisa Richardson, CDC Division of Cancer Prevention and Control; and Saul Weingart, Rhode Island Hospital and Brown University.
REVIEWERS: To ensure that it meets institutional standards for quality and objectivity, this Proceedings of a Workshop—in Brief was reviewed by Elizabeth A. Sarma, National Cancer Institute, and Electra D. Paskett, Ohio State University. Leslie J. Sim, National Academies of Sciences, Engineering, and Medicine, served as the review coordinator.
STAFF: Jennifer Lalitha Flaubert, Ruth Cooper, Anesia Wilks, and Sharyl Nass, Board on Health Care Services, Health and Medicine Division, National Academies of Sciences, Engineering, and Medicine
SPONSORS: This workshop was supported by the Gordon and Betty Moore Foundation.
For additional information regarding the workshop, visit https://www.nationalacademies.org/event/10-06-2021/achieving-excellence-in-cancer-diagnosis-a-workshop.
Suggested citation: National Academies of Sciences, Engineering, and Medicine. 2022. Achieving excellence in cancer diagnosis: Proceedings of a workshop—in brief. Washington, DC: The National Academies Press. https://doi.org/10.17226/26505.
Health and Medicine Division
Copyright 2022 by the National Academy of Sciences. All rights reserved.