Advancing Global Cancer Symptom Science: Insights and Strategies From the Inaugural Cancer Symptom Science Expert Meeting

Abstract

BACKGROUND

The inaugural “Cancer Symptom Science Expert Meeting,” held in Lausanne, Switzerland, on October 11–12, 2023, brought together 40 nurse scientists from seven countries. The event aimed to enhance collaboration across the global symptom science community; identify common research interests, gaps in knowledge, and opportunities for research; and develop strategies to address challenges and accelerate symptom science research internationally.

OBJECTIVE

The aim of this white paper is to summarize the discussions and recommendations deliberated during the meeting and introduce the Global Research Alliance in Symptom Science (GRASS).

METHODS

This two-day meeting featured presentations that highlighted critical issues and unanswered questions in cancer symptom science and other chronic conditions. Attendees identified four core topic areas based on the knowledge gaps reflected throughout the presentations. Of note, the co-occurrence of cancer with other chronic conditions (e.g., cardiovascular disease, diabetes) that may share similar contributors and underlying mechanisms were included in broader context. Four working groups (WGs) were formed to identify gaps and opportunities associated with each topic and to outline strategic directions and essential actions to advance symptom science.

RESULTS

The WGs developed recommendations on four core topic areas. WG1 explored optimal approaches to collect, analyze, and use symptom data for research and clinical purposes. WG2 addressed the development of a minimum dataset or common data model for symptom science research. WG3 focused on enhancement of best practices in implementation science strategies to improve uptake of evidence-based symptom management strategies in routine clinical care. WG4 addressed the questions of capacity building and infrastructure for the creation of a global alliance in symptom science (GRASS).

DISCUSSION

WGs’ recommendations underscore the commitment of an international coalition of scientists to advance symptom science. The symposium established the groundwork for the group to constitute GRASS, a global research alliance dedicated to symptom science in cancer and other chronic conditions. Future directions include establishing regular scientific meetings, fostering interprofessional collaboration, and engaging with symptom scientists.

Over the past several decades, oncology nurse scientists established the foundational elements of research in symptom science. Based on seminal work (Dodd et al., 2011; Dodd, Janson, et al., 2001; Dodd, Miaskowski, et al., 2001; Ferrell, 1995; Given et al., 2001, 2008; McCorkle, 1987; McCorkle & Young, 1978; Miaskowski, 2016; Miaskowski et al., 2015, 2017) that described the detrimental effects of unrelieved symptoms on patients’ physical (Torstveit et al., 2021), psychological (Suskin et al., 2022; Van Onselen et al., 2010), and cognitive (Torstveit et al., 2021) functioning, as well as on quality-of-life (QOL) outcomes (Jaffee et al., 2017; Nho et al., 2017; Pud et al., 2008) and overall survival (Kerrigan et al., 2020), an evaluation of symptoms and their effects is a priority for research and clinical practice. While large gaps in knowledge of patients’ symptom experiences across ages throughout the continuum of cancer care exist, some foundational elements are notable:

• ■ Patients across the age span experience an average of 10 to 14 co-occurring symptoms during and following cancer treatment (Hinds et al., 2021; Miaskowski, 2016; Miaskowski et al., 2015, 2017).

• ■ Co-occurring symptoms can aggregate into clusters. Symptom clusters are defined as two or more concurrent symptoms that form a stable symptom group, which may have common underlying mechanisms and shared outcomes, and which have a temporal dimension (Hockenberry et al., 2010; Hooke & Linder, 2019; Miaskowski et al., 2017; Rogers et al., 2019).

• ■ A large amount of inter-individual variability exists in patients’ symptom experiences across the age span (Hockenberry et al., 2010; Hooke & Linder, 2019; Miaskowski, 2016; Miaskowski et al., 2015, 2017; Rogers et al., 2019).

• ■ Across various types of cancer and treatments, a subgroup of patients with a high symptom burden can be identified (Buckner et al., 2014; Cheng et al., 2023; Hinds et al., 2021; Miaskowski, 2016; Miaskowski et al., 2015, 2017; Rogers et al., 2019; Withycombe, Eldridge, et al., 2022; Withycombe, McFatrich, et al., 2022).

• ■ Progress has been made in the development and testing of interventions to manage single and co-occurring symptoms, as well as symptom surveillance (Berry et al., 2015; Mooney et al., 2024; Nakamura et al., 2024).

These consistent findings emerged over decades of research and established the foundation for the field of symptom science. As highlighted in Figure 1, research led by nurse scientists was woven throughout the fabric of scientific progress in cancer care since the signing of the U.S. National Cancer Act in 1971 (Tiwari & Roy, 2012) and ongoing symptom science discoveries within nursing (Hinds et al., 2021; Miaskowski et al., 2017) have occurred over the past five decades. With this history as a foundation, two events were held in 2023 that fostered the development of global cancer symptom science initiatives. First, in April, a Precision Health Symptom Science Mini Symposium supported by the Oncology Nursing Society (ONS) was held just prior to the annual ONS Congress^® in San Antonio, Texas. Based on the Mini Symposium feedback and evaluations from national and international attendees, a focused expert meeting was planned. This second event, an international Symptom Science Expert Meeting, convened in October 2023 in Lausanne, Switzerland. These meetings culminated in the formation of the Global Research Alliance in Symptom Science (GRASS). GRASS’s processes, infrastructure, mission, and vision are described in this white paper.

FIGURE 1.

Historical Overview of Oncology Nurse–Led Symptom Science Initiatives

* Oncology Nurse Scientists recognized by NCI Division of Cancer Prevention in celebration of the 50th Anniversary of the U.S. National Cancer Act.

FIRE—Fatigue Initiative through Research and Education; NCAT— National Center for Advancing Translational Sciences; NCI—National Cancer Institute; NINR—National Institute of Nursing Research; ONS—Oncology Nursing Society

Global Symptom Science Need

Research on symptoms constitutes a primary area of interest for nurse scientists. The broad aims of this research are to increase knowledge of patients’ symptom experiences and develop, test, and implement optimal symptom management interventions. A variety of approaches are used to achieve these aims, including descriptive studies of symptom manifestations in various disease conditions; an evaluation of interventions for symptom prevention and treatment; and studies to determine symptoms’ underlying mechanisms. An overall objective is to identify individuals at increased risk for the highest symptom burden and to develop targeted interventions to prevent or lessen symptom severity. This approach includes a comprehensive evaluation of patients’ demographic and clinical characteristics, as well as behavioral, social, and environmental determinants of health. Equally important, underlying mechanisms need to be characterized to guide the development and testing of tailored symptom management interventions (Miaskowski et al., 2017).

As the healthcare paradigm evolves, alignment of symptom science research within the framework of precision health, including whole person care (National Center for Complementary and Integrative Health, 2021), is an imperative. In order to achieve a comprehensive understanding of common symptoms across diverse populations and age groups, large studies are needed that provide sufficient statistical power to ensure the reliability and generalizability of the findings. The establishment of collaborations among researchers from diverse academic and practice institutions and disciplines is needed to achieve these goals (Dorsey et al., 2019).

Process

This collaborative spirit is exemplified by two recent initiatives described previously. First, the Precision Health Symptom Science Mini Symposium included a number of key themes (i.e., symptom management and QOL, technology and implementation, data science and biomarkers, health behaviors and diversity across populations and regions, and research models) and provided a comprehensive exploration of contemporary topics in symptom science. Second, the inaugural international “Cancer Symptom Science Expert Meeting,” held in Lausanne, Switzerland, brought together 40 nurse scientists from seven countries (i.e., Canada, Denmark, Netherlands, Spain, Switzerland, United Kingdom, and United States). This two-day meeting aimed to enhance collaboration within the symptom science community; identify common research interests and knowledge gaps; and develop strategies to mitigate challenges and accelerate international symptom science research. Day 1 featured presentations that highlighted critical issues and unanswered questions in cancer symptom science. On the second day, attendees identified four core topic areas based on the knowledge gaps identified on day 1. Four working groups (WGs) were formed to identify opportunities and challenges associated with each topic and to outline strategic directions and essential actions to advance symptom science. In the following, we summarize the discussions and recommendations addressed in each of the WGs and introduce GRASS.

WG1: Optimal Approaches to Collect, Analyze, and Use Symptom Data for Clinical and Research Purposes

Data are a central component of research and clinical practice and serve as a core aspect of how cancer symptoms are understood, evaluated, and managed. In research, these data are harnessed to explore the epidemiology of cancer symptoms, investigate underlying mechanisms, identify symptom clusters and profiles, and evaluate the effectiveness of strategies for symptom assessment and management. In clinical practice, symptom data are used routinely to identify disease-related symptoms, characterize treatment toxicities, tailor symptom management, and evaluate intervention effectiveness. WG1’s deliberations focused on an exploration of optimal approaches to collect, analyze, and use symptom data to advance symptom science.

Researchers and clinicians use diverse data sources to develop a holistic understanding of patients’ symptom experiences. Symptom data are collected through patients’ self-reports, clinical observations and history taking, wearable devices, biologic specimens, and assessments of patients’ behavioral, environmental, and social contexts. Approaches to symptom data collection are rapidly evolving. For example, efforts are underway to standardize symptom assessments in adult (Basch et al., 2014; Reeve et al., 2014) and pediatric (Roth et al., 2023) populations. Patient-reported outcomes (PROs) in adult populations are being integrated into electronic health records (EHRs) (Hassett, Cronin, et al., 2022). Similarly, nurse-led efforts to establish evidence-based symptom assessments are now provided by the Nursing Discipline Committee in the Children’s Oncology Group (Withycombe, Eldridge, et al., 2022; Withycombe, McFatrich, et al., 2022). In addition, a nurse-led multisite study and subsequent larger-scale multisite study (Hinds et al., 2013, 2019; Reeve, McFatrich, Mack, Maurer, et al., 2020; Reeve, McFatrich, Mack, Pinheiro, et al., 2020) validated the Patient-Reported Outcomes Measurement System (PROMIS^©) Pediatric Measures. Furthermore, a nurse-led interprofessional team developed, validated, and scaled for worldwide implementation of the Patient-Reported Outcomes version of the Common Terminology Criteria for Adverse Events (PRO-CTCAE^®) Measurement System and the PRO-CTCAE Pediatric Module (Jacobs et al., 2022; Kluetz et al., 2016; Minasian et al., 2022; National Cancer Institute, 2025; Reeve et al., 2017; Reeve, McFatrich, Mack, Maurer, et al., 2020; Reeve, McFatrich, Mack, Pinheiro, et al., 2020). Across age groups, efforts are ongoing to standardize reports from patients and their family caregivers in clinical trials (Roth et al., 2023). In addition, the potential for wearable devices to inform symptom management is being explored (Wright et al., 2018). Finally, studies of remote symptom monitoring systems (Berry et al., 2014; Harris et al., 2018; Hassett, Cronin, et al., 2022; Hassett, Wong, et al., 2022; Maguire et al., 2021) demonstrate their potential to promote symptom control and improve outcomes in patients with cancer.

Both de novo and a priori analytic approaches can be used to characterize symptoms in patients with chronic conditions (Miaskowski et al., 2007). For symptom clusters, statistical techniques, such as exploratory and/or confirmatory factor analysis (Mack et al., 2020; Pozzar et al., 2022; Thornton et al., 2023), are used for de novo identification of symptom clusters. For identified symptoms/symptom clusters, research focused on a priori symptom clusters (Hammer et al., 2022; Harris, Dodd, et al., 2022; Harris, Kober, et al., 2022) is used to identify subgroups of patients with distinct symptom profile dimensions. Specifically, these subgroups of patients can be identified using symptom dimension scores (e.g., occurrence, severity, distress, interference) (Papachristou et al., 2019); hierarchical cluster analysis (Dodd, Janson, et al., 2001); and latent variable modeling (Buckner et al., 2014; Cheng et al., 2023; Harris, Dodd, et al., 2022; Harris, Kober, et al., 2022; Hinds et al., 2021; Miaskowski et al., 2017; Wang et al., 2023; Weaver et al., 2022) (e.g., latent class analysis for categorical data [e.g., symptom occurrence] and latent profile analysis for continuous data [e.g., symptom severity]). Recent advances in symptom clusters research include the use of network analysis to assess associations between and among symptoms and symptom clusters; predictive modeling procedures to analyze “big data”; and natural language processing to analyze unstructured text from the EHR (Harris, Dodd, et al., 2022).

Challenges and Gaps

While symptom data collected during routine clinical care is often used in research, these data are currently underutilized. Harnessing clinical data for research purposes has the potential to accelerate advancements in symptom science. However, scientists face several barriers to acquire and use clinical data. First, scientists without a clinical role may contend with gatekeeping during recruitment (Bell et al., 2020). An additional hurdle is the unstructured and unstandardized nature of clinical data. For example, the narrative text of clinical notes may include symptom data that are not captured systematically in the EHR. Moreover, scientists need to organize and structure the integration of diverse data types, such as symptom scores, biomarker levels, images, or text, into one dataset. A second challenge is the sheer volume of clinical data and the speed at which it accumulates. EHRs include a wide variety of data types that are updated during each patient contact. Specialized statistical and analytical techniques (e.g., machine learning) are required to analyze and interpret the “big” and unstructured data collected during clinical care (Harris et al., 2023). Third, a distinct challenge in pediatric oncology is ensuring that the voices of patients, caregivers, and clinicians are considered in symptom assessments (Hinds et al., 2017). This underscores the importance of understanding congruence in symptom reporting among these three groups. Finally, the homogeneity among patients enrolled in research studies may limit the generalizability and applicability of research findings (Harris, Dodd, et al., 2022).

Equally important, data collected for research purposes may not be readily interpretable or accessible to clinicians. Systems that collect, analyze, and interpret symptom scores at the point of care are uncommon. More frequent symptom assessments are associated with more timely identification of severe symptoms (Innominato et al., 2021). However, limited interoperability between remote symptom monitoring applications and EHR systems constrains clinicians’ ability to identify and respond to severe symptoms between visits. To date, remote symptom monitoring has been implemented predominantly for research purposes at specialized cancer centers (Hassett, Cronin, et al., 2022). In the future, institutions without the infrastructure or staff to support the implementation of these systems may be at a disadvantage that would lead to further disadvantages if these systems became reimbursed because they improve patient outcomes.

In addition, several scientific questions warrant consideration. First, it is crucial to determine the optimal approach to symptom assessment, as scientists must choose from various valid and reliable symptom measures, each with its own recall period (Bull et al., 2022). The heterogeneity of these measures limits comparisons across studies and analyses of pooled data, making it equally important to establish the most appropriate recall period for symptom reporting. While the U.S. Food and Drug Administration (Norquist et al., 2012), the European Medicines Agency (2016), and organizations like the International Society for Quality of Life Research (Reeve et al., 2013) provide guidance, they generally do not specify exact recall periods. Instead, they emphasize the importance of justifying the chosen recall period based on the specific context of the study, the nature of the disease, and the concept being measured. Shorter recall periods (e.g., seven days or less) enhance the accuracy and reliability of reported symptoms; align with regulatory expectations; and ensure that the data are a faithful representation of a patient’s current health status (European Medicines Agency, 2016; Norquist et al., 2012; Reeve et al., 2013). Second, the optimal strategies to promote engagement and mitigate response bias in remote symptom monitoring systems are unclear. In a scoping review of patient engagement in eHealth (i.e., technology-based) interventions (Oakley-Girvan et al., 2022), engagement was found to decline over time, often most precipitously after the first week of use. Equally important, the role of caregiver support, treatment fidelity, self-management behaviors (e.g., medication/activity adherence [Shi et al., 2023]), and patient values in shaping symptom trajectories needs to be determined. An additional challenge for pediatric oncology is directly related to the rare disease status (few children diagnosed with cancer) and small sample sizes, particularly in early phase trials (Hinds et al., 2017).

Opportunities

Despite these challenges, several opportunities exist to advance symptom science. The availability of large datasets (e.g., UK Biobank, All of Us, Surveillance, Epidemiology, and End Results–Medicare Health Outcomes Survey) offers a unique opportunity to assess symptoms on a broader scale. The development of a standardized nomenclature for EHR data is pivotal because it will facilitate streamlined data extraction for symptom research. Real-time data analysis strategies need to be developed and tested to be able to interpret and act on symptom data at the point of care. The ability to analyze EHR data for changes in symptoms over time will provide valuable insights into the dynamic nature of symptoms and the opportunity to preemptively treat them. Promising avenues for research include symptom monitoring systems and the creation of personalized symptom management plans. In addition, utilizing real-time data to evaluate the effectiveness of various symptom management interventions will inform the implementation of evidence-based guidelines as well as provide an opportunity for comparative effectiveness research. Collaboration with patients, family caregivers, and clinicians is essential to enhance the design and implementation of evidence-based symptom management interventions. In addition, invested partners need to advocate for symptom management data to be incorporated into the EHR and into clinical decision-making. These efforts will transform data-driven approaches to prevent and manage symptoms. Opportunities and strategies to optimize the collection, analysis, and identification of symptom data are listed in Table 1.

TABLE 1.

Challenges, Opportunities, and Strategies to Optimize Data Collection, Analysis, and Use of Symptom Data

Challenges	Opportunities	Strategies
Data collection
Limited engagement among invested partners	Enhance invested partner awareness and collaboration: foster a culture of engagement, collaboration, and shared understanding among invested partners in symptom science research.	■ Educate clinicians and administrators about the purpose and potential of symptom science research. ■ Develop and fund interprofessional training programs in symptom science research. ■ Conduct research that identifies factors that promote patient, caregiver, and clinician engagement in symptom assessment and management.
Clinician gatekeeping during recruitment	Foster clinician support and participation: create a supportive environment for recruitment and increase clinician involvement in research.	■ Build partnerships with clinicians by including them as co-investigators on grants, protocols, and manuscripts. ■ Educate colleagues about the impact of symptom science to improve patient outcomes.
Lack of best practices in routine data collection	Establish standardized measurement protocols: ensure consistency, accuracy, and comparability of symptom data through a common data model.	■ Develop a standardized nomenclature for symptom documentation in the electronic health record (EHR). ■ Develop a minimum dataset or common data models for systematic and standardized data collection. ■ Use valid and reliable measures with a recall period of seven days or less.
Limited racial, ethnic, and geographic diversity among participants in symptom science studies	Promote diversity and inclusivity: ensure that symptom science research represents and addresses the needs of diverse populations.	■ Recruit and educate a more diverse research workforce. ■ Standardize data collection related to race, ethnicity, gender identity, socioeconomic status, educational attainment, and other social and structural determinants of health. ■ Build partnerships with people and organizations with knowledge of and access to underrepresented groups of patients. ■ Collaborate with invested partners to co-design symptom management interventions and identify symptom science research priorities.
Analysis
Complex statistical analyses are needed to answer complex questions in symptom science.	Utilize advanced analytic techniques to gain a comprehensive understanding of symptoms.	■ Use exploratory factor analysis to identify symptom clusters de novo. ■ When sample sizes are sufficient and relevant expertise is available, use network analysis to identify sentinel symptoms. ■ Use latent variable modeling procedures to identify subgroups of patients with distinct symptom profiles. ■ Develop and test machine learning models to analyze large quantities of structured and unstructured data (e.g., narrative clinical notes).
Use in clinical practice
Limited interoperability between remote symptom monitoring systems and EHRs Limited interpretability of research data	Develop strategies to enhance interoperability and interpretability of research data for tailored and timely interventions.	■ Collaborate with informaticians to develop and test strategies that integrate remote symptom monitoring applications into the EHR. ■ Develop and test algorithms that use findings from the above analytical approaches to identify patients at increased risk for adverse symptom experiences at the point of care. ■ Conduct research with patients, family members, and clinicians to understand what each invested partner wants included in a symptom report.

WG2: Development of a Minimum Dataset or Common Data Model for Symptom Science Research

Over the past 25 years, support for the identification and utilization of common data elements (CDEs) has grown worldwide to be able to compare findings across studies and populations and to strengthen findings by combining datasets. A CDE is defined by the National Institutes of Health as a “standardized, precisely defined question, paired with a set of allowable responses, used systematically across different sites, studies, or clinical trials to ensure consistent data collection” (National Library of Medicine, n.d.).

Between 2015 and 2020, a series of landmark papers were published that described approaches to increase the use of CDEs in symptom science research (Cohen et al., 2015; Corwin et al., 2017; Hickey et al., 2019; Moore et al., 2016; Page et al., 2018; Redeker et al., 2015). The first of these landmark papers described the best practices for choosing, utilizing, managing, and sharing CDEs within and across institutions. Recommendations included minimum sets of sociodemographic (i.e., age, birth sex, gender identity, culture identity, education, socioeconomic status, and occupation) and symptom (i.e., pain, fatigue, sleep disturbance, anxiety, depression, and cognitive disturbance) CDEs (Redeker et al., 2015). The second paper recommended a set of CDEs for self-management science (Moore et al., 2016). The third paper recommended a minimum set of biomarker CDEs for symptom and self-management science (Page et al., 2018). The fourth paper in the series highlighted the utility of CDEs to deal with data integration, particularly in relationship to precision health initiatives (Hickey et al., 2019). As a result of this seminal literature, coupled with support from professional organizations such as ONS (Von Ah et al., 2019) and the American Society of Clinical Oncology (Osterman et al., 2020), researchers made strides to incorporate recommended CDEs into data-based studies (Breazeale et al., 2022; Chen et al., 2022; Kamp et al., 2023; Kim et al., 2020; Lerret et al., 2022; McClintock et al., 2023; Sohl, Befus, et al., 2023; Sohl, Sadasivam, et al., 2023; Topaz et al., 2021; Wojeck et al., 2021); develop new CDEs (Edvardsson et al., 2019; McGilton et al., 2020); and build repositories to facilitate data sharing (Moore et al., 2020; Perazzo et al., 2019).

Challenges and Gaps

Despite this progress, several challenges have hindered the incorporation of CDEs into symptom science research. For example, researchers at one institution reported significant challenges with adopting CDEs for use across pilot studies of self-management interventions for chronic conditions (Corwin et al., 2017; Moore et al., 2020). Key issues included delays in reaching consensus on CDEs for demographic data; limitations in the application of CDEs across diverse patient groups or health conditions; and resistance among investigators to adopt established CDEs into their research programs (Corwin et al., 2017; Moore et al., 2020). Notably, these challenges are magnified when CDEs are taken to a global scale. For example, pertinent demographic data vary within (e.g., rural versus urban) and across (e.g., level of education) countries. In addition, CDEs may not be translated or validated across various languages or cultures (Tao et al., 2023). In addition, differences in social constructs across populations warrant consideration. For example, certain data (e.g., demographic) may be culturally inappropriate to collect in some populations. Therefore, combining datasets across populations may require novel analytic techniques.

In addition, multiple gaps remain in how to incorporate new CDEs in symptom research on a global scale. While three of the foundational papers recommended initial CDEs for symptom and/or self-management science (Moore et al., 2016; Page et al., 2018; Redeker et al., 2015), adoption of new CDEs has not occurred. This lack of evolution may be due to a variety of factors including decreased prioritization of CDE development and lack of national and international leadership on this initiative. Another barrier is the growing number of institutional initiatives that have interprofessional teams developing large-scale platforms to collect symptom data from all patients within their systems. One example is the EHR-integrated, PRO-based electronic symptom management system that uses the PRO-CTCAE instrument to capture symptom data (Hassett, Cronin, et al., 2022; Hassett, Wong, et al., 2022). As patient data accrue, adopting new CDEs outside of the PRO-CTCAE question bank may not occur. As a result of these issues, opportunities to combine datasets across institutions for large-scale data science studies are limited.

Another key issue underlying the delays in adoption of CDEs is that they are inherently specific and rigid. This lack of elasticity restricts the evolution of science. For example, while the initial symptom CDEs that were recommended in the aforementioned landmark papers included the PROMIS^© Applied Cognition v1.0 and General Concerns v1.0 instruments to measure cognitive disturbance (Redeker et al., 2015), these instruments were revised and the use of the previous versions is no longer recommended by PROMIS (Hassett, Cronin, et al., 2022; HealthMeasures, 2022; Lai et al., 2014). In addition, given that each of the symptom CDEs that were initially recommended contain two or more short form versions, merging datasets that used different versions is challenging. However, methods exist to pool symptom scores from different study instruments (Atherton et al., 2015). Furthermore, while the recommended biomarker CDEs were comprehensive and up to date at the time of publication, biomarker development is advancing at an exponential rate. For example, evidence suggests that perturbations in the microbiome and epigenome affect a variety of symptoms (e.g., cognitive disturbance, depression) and other outcomes in patients with various chronic conditions (Ray et al., 2022; Song & Bai, 2021). Given the significant costs associated with biomarker research, the inclusion of new biomarkers as part of CDEs warrants careful consideration.

Opportunities

Given these challenges, we believe it is most feasible to propose a common data model (CDM) for symptom science as an alternative to CDEs. CDMs are comprised of broad categories of data (i.e., domains) that capture common concepts across studies or institutions and are used to “standardize and facilitate the exchange, pooling, sharing, or storing of data from multiple sources” (Garza et al., 2016, p. 333). CDMs are used to reformat data into a standardized structure and coding format that allows for data to be combined for downstream analyses. Examples include the Observational Medical Outcomes Partnership CDM (Observational Health Data Sciences and Informatics, 2025) and Patient-Centered Outcomes Research Network (PCORnet, 2023) CDM. Use of a CDM allows for flexibility in the inclusion of additional data elements and concepts to support the evolving nature of research (e.g., demographic characteristics, biomarkers). This approach could have greater feasibility and acceptability among invested partners and improve the willingness of researchers to share data, thereby providing a framework for the initial development of a central data repository.

The Symptom Science CDM will contain domains that encompass specific concepts that are central to symptom science. Each primary domain will contain data elements that measure the specific domain concept in different ways. For example, the “Clinical Factors” domain will include recommendations for collecting data on patients’ clinical condition(s), time since diagnosis, and acute and chronic treatments (past and present). Data elements contained within each domain will be listed using a tiered system, starting with “strongly recommended” data elements for collection followed by secondary and tertiary data elements. While specific symptom measures or instruments may be listed within each domain, the model will not endorse specific CDEs, but similar underlying constructs. Therefore, this initial CDM may be viewed as a catalog of data elements with each domain containing one or more operational definitions for the concept being measured. Over time, the model may endorse specific CDEs.

To inform the structure and composition of the Symptom Science CDM, this WG will review existing CDMs to assess model structure and evaluate the effectiveness of each model for their specific use. Next, a review of current recommendations for CDEs and/or minimum datasets from national and international professional organizations will be conducted to identify core domains for inclusion in the CDM. To increase communication, collaboration, and data sharing among symptom scientists around the globe, a system for combining data will be established. In addition, a system for storing biologic data (e.g., genetic, microbiome) will be developed.

WG3: Enhancement of Best Practices in Implementation Science

Routine monitoring of PROs that capture the effects of patients’ symptom experiences on the “whole person” is a priority for healthcare systems worldwide (Black et al., 2016; Howell et al., 2013) and an integral component of value-based health care (Silveira Bianchim et al., 2023). Evidence from randomized controlled trials suggests that symptom monitoring using PROs reduces symptom and emotional distress, decreases healthcare utilization (e.g., hospitalization, avoidable emergency visits), and improves person-centered communication, treatment adherence, QOL (Licqurish et al., 2019; Lizán et al., 2021; Wang et al., 2023), and survival (Basch et al., 2017). However, most of these effects are reported from randomized controlled trials rather than from PRO monitoring as part of routine clinical care. Of note, a seminal review conducted by Kotronoulas et al. (2014) identified that routine PRO use was associated with increased patient outcome discussions with clinicians. In addition, better symptom management, supportive care, and patient satisfaction were reported across a number of studies (Kotronoulas et al., 2014). Findings across other studies were mixed. For example, several pragmatic trials evaluated in a comprehensive review (Zigler et al., 2024) examined PRO symptom surveillance and guideline-based management with comparisons to usual care. The findings between studies were mixed due to inconsistencies and challenges across health systems. Routine symptom monitoring will not translate into better outcomes unless clinicians and patients act on these data. The clinical utility of these data requires: (a) its integration within the clinical encounter for person-centered communication; (b) use of data to inform clinical decision-making and tailoring of symptom management interventions (Howell & Liu, 2011); and (c) intensification of interventions and follow-up for patients with the highest symptom burden (Astrup et al., 2017). Timely assessments and interventions are needed to decrease symptom burden and optimize the plans of care (Rosenzweig & Mazanec, 2023).

Challenges and Gaps

Despite the dramatic increase in the use of PROs for symptom assessment, efforts to adopt this approach as the standard of care in routine clinical practice have had mixed success (Lutz et al., 2022). Numerous clinical barriers exist to the use of real-time PRO data including lack of knowledge among clinicians about how to interpret PRO data and changes in data over time; how to use these data in clinical decision-making and treatment planning; and actionability of some symptoms (e.g., fatigue). Additional clinical barriers include disruption in workflow; longer visit times; lack of a technology infrastructure to facilitate data collection; and poor engagement of end users in selecting PROs that matter and are relevant based on the patients’ clinical condition (Foster et al., 2018; Nguyen et al., 2021). Patient barriers include time to complete PROs; PROs available in a limited number of languages; poor digital literacy; relevance of PROs; and their perceived lack of value and use in clinical care. The integration of PROs for symptom management requires systemic changes and considerable efforts to reduce these barriers within organizations to accelerate their uptake and use across the continuum of patient care (Cheung et al., 2022; Li et al., 2016).

Implementation science is a field of research that seeks to improve how evidence-based interventions and health service innovations such as PROs are successfully adopted, implemented, and maintained in healthcare delivery systems (Estabrooks et al., 2018). However, most symptom science frameworks do not incorporate implementation science as a core element (Saligan, 2019). Achieving the integration of PROs into clinical care will require that clinical, health service, and population health researchers work with clinicians, administrators, patients, and family members to design studies that are clinically feasible and informative (Stover et al., 2021).

Opportunities

Several opportunities exist to improve the implementation of PROs in routine cancer care:

1. Synthesize literature on implementation methods/strategies effective for promoting/sustaining use of PROs for symptom management and evaluate effects: While review papers (Lyu et al., 2024) are emerging that evaluate implementation methods for PRO uptake, the quality of implementation, including the use of implementation theory, the lack of tailoring of implementation strategies to barriers, and the facilitation of change in practice behaviors that can sustain uptake in routine care, remains poor (Mazariego et al., 2022). Additional scoping or systematic reviews of the literature may provide further insights into what implementation strategies are effective in specific healthcare contexts. Standardization in the application and reporting of implementation theories, implementation approach and strategies, and best practices to educate practice change facilitators are needed (Proctor et al., 2011). Quality indicators that foster accountability for symptom management need to be developed for accreditation standards and for Magnet^® recognition. In addition, best practices in service design and transformed models of care that mobilize nurses as first responders to PRO data require further development, testing, and ultimately scaling up to reach all populations (Graupner et al., 2021).

2. Evaluate and adapt new implementation strategies that target patient use of PRO data and patient activation: Implementation of PROs has focused mainly on uptake by clinicians. A need exists to design and test implementation strategies that promote patient involvement in the selection of PROs; how to disseminate these data to patients (e.g., graphs, box plots, change over time); how to interpret data and change in data over time; and education on how to use these data for symptom self-monitoring and self-management. Early engagement of patients in the selection of PROs is recommended and best practices must be synthesized (Carlton et al., 2020). Targeted education of patients about the use of PROs for symptom self-management is crucial (Lavallee et al., 2016). Minimum datasets for tracking patient activation and engagement in symptom self-management need to be integrated into PRO programs (Moore et al., 2016). Digital technology could be used for automation of standardized symptom self-management information. In addition, briefing documents or position statements could be disseminated to key influencers (e.g., patient advocacy groups, health policy leaders, payers) that call for better symptom care and coverage of the expenses associated with including PRO data in clinical care.

3. Education of clinicians in use of PROs and implementation science: Clinicians can benefit from education on how to deliver effective self-management support that will activate patients to be co-collaborators in achieving better symptom outcomes (Howell et al., 2021). In addition, clinicians need more education on how to use PROs for person-centered communication, clinical decision-making and treatment planning, and provision of self-management support (Skovlund et al., 2020). Clinician education should include how to score and interpret changes in PRO data over time; how to integrate PRO data in person-centered communication; how to engage patients to promote behavior change; and how to use these data in algorithm-based precision symptom science (Moore et al., 2018). Consensus on an implementation science core curriculum and global access to implementation science education programs are needed (Chambers et al., 2017; Padek et al., 2018). Ideally, clinical symptom monitoring specialist roles can be implemented across institutions. Of note, among institutions that have nurse navigators, symptom monitoring is often included in this role (Rocque et al., 2022). Furthermore, the use of mHealth tools to support symptom monitoring is being developed and tested across diseases (Maguire et al., 2021; Rudin et al., 2017; Schmidt et al., 2022).

WG4: Creation of a Global Research Alliance in Symptom Science (GRASS)

While WGs1–3 focused on content gaps in symptom science, WG4 was tasked with evaluating the need to formalize a network to move symptom science forward across populations.

Challenges and Gaps

WG4’s assessment was based on the expert meeting’s day 1 presentations that included the history, state of the science, and gaps in symptom science; coupled with robust discussions among the experts. A major challenge was identified. Specifically, the field of symptom science lacks continuity in a vision, a mission, and objectives. The experts determined that collaboration across continents could fill gaps in symptom science and develop a precision health approach to care that can be adapted throughout populations.

Opportunities

Based on the evaluation of materials and needs, WG4 recommended the formation, naming, and implementation of an international network in symptom science. Meeting attendees unanimously agreed with this recommendation. The organization name of Global Research Alliance in Symptom Science (GRASS) emerged as both a specific reflection of the network’s purpose and the vision of growing into the future, as the acronym implies.

In addition to assessing the need, WG4 provided recommendations for developing the GRASS infrastructure. The proposed infrastructure was for GRASS to become a collaborative alliance of scientists, learners, and care recipients from a variety of academic, clinical, and community entities that conduct research, implement evidence-based interventions, and educate a broad constituency (e.g., patients, consumers, insurers, policymakers, clinicians) on symptom science. An “alliance” was selected to formalize the network due to the goals and type of structure that seemed most appropriate. Specifically, following the “Four C’s” of complementarity, congruent goals, compatibility, and change (Coff, 1995), such an alliance can accomplish strategic goals in global symptom science and its translation to clinical practice. The alliance is facilitated by nurse scientists active in the field of symptom science. Interprofessional members/collaborators are essential to bring a variety of needed perspectives.

GRASS is committed to conducting cutting-edge and innovative research and facilitating the implementation of research findings into practice. This interprofessional alliance is envisioned to grow to include experts from nursing, psychology, medicine, biology, social work, medical engineering, biostatistics, patient and public sectors, data science, and digital health. GRASS will assess the scientific landscape and prioritize the development and execution of projects within the different task forces outlined previously.

Of note, while GRASS was established seamlessly through unanimous agreement as an outcome of the expert meeting, implementing goals internationally will not be without some challenges. International research can be difficult to fund across sites; may face barriers with differing regulatory requirements; may have challenges in translating study materials into different languages; and may face obstacles in coordinating the collection of samples for biomarker data, to name a few (Alberti et al., 2024). One of the goals of this alliance is to address these challenges and create streamlined protocols.

To promote collaborative research and continuing education, meetings will be held on a regular basis. This approach will provide support to and mentorship of the next generations of researchers in the field of precision health and symptom science. Leaders in the field will present the current state of the science in research and its translation to practice, as well as directions for future research. Attendees will have the opportunity to present projects at GRASS meetings and receive feedback to strengthen their projects.

In alignment with the previously mentioned activities and meetings, the intention of GRASS efforts is to stay aligned with major healthcare professional organizations and their guidelines and policies across countries. It is anticipated that some GRASS-related work will inform these organizations’ guidelines related to clinical research and practice related to precision health symptom science. Once further established, it is anticipated that GRASS will become a resource for professional and public sectors.

Discussion

Founded on the history of oncology nurse–led symptom science (see Figure 1), GRASS was established to build on this legacy with the objective of reducing the global burden of symptoms across multiple single and co-occurring chronic conditions. Two key events in 2023 (i.e., ONS-affiliated Mini-Symptom Science Symposium in the United States, and Symptom Science Expert Meeting in Switzerland) provided the infrastructure for the development of GRASS. The four WGs outlined the challenges, opportunities, and strategies to move symptom science forward. Importantly, as a global alliance, the involvement of diverse populations, including those who are underserved and previously underrepresented, will be placed at the center of these efforts.

Key highlights from the WG summaries with added details are provided in Table 2. In brief, there is a need to develop a more uniform system for symptom data collection across populations and age groups. In alignment, the development of CDMs can aid in combining symptom data for the conduction of large data science studies. Conducting large studies on comprehensive datasets can expedite filling essential gaps in knowledge (e.g., identifying symptom risk profiles; developing effective, person-centered symptom management strategies). Novel methods such as the use of artificial/augmented intelligence and machine learning techniques can aid in expediting answers to key research questions. Equally important, incorporating a model of implementation science can guide symptom assessment and management across different types of practices, communities, and populations. Finally, the expert meeting attendees concluded that there is a need to develop a global alliance to move this work forward.

TABLE 2.

Summary of WGs

	WG1 Optimizing approaches to collect, analyze, and use symptom data for clinical and research purposes	WG2 Developing a minimum dataset for symptom science research	WG3 Enhancing best practices in implementation science to improve uptake of evidence-based symptom management interventions	WG4 Creating a Global Research Alliance in Symptom Science (GRASS)
Challenges	Symptom data are undercollected and underutilized in research and clinical practice; pediatric patients may lack a voice in symptom assessments.	Lack of consistent collection of symptom data across research and clinical practice and across institutions	Lack of symptom- and treatment toxicity–focused PROs as a standard of care	Lack of shared vision, mission, and objectives regarding the evolution of symptom science
Opportunities	Collect multiple forms of symptom data including self-report, clinical observations, wearable devices, biologic specimens, and behavioral, environmental, and social contexts; strengthen efforts underway to collect symptom assessments from patients across age groups, family caregivers, and clinicians.	Develop minimum datasets using common data models to capture similar constructs of symptom experiences across studies.	Incorporate a model of implementation science to use PROs for assessment and to guide symptom management strategies.	Develop an alliance to drive global symptom science.
Strategies	Establish collaborative partnerships; develop standardized methods for symptom data collection.	As a minimum dataset is developed, translate instruments into different languages across cultures to enhance uniformity; develop a system for combining datasets across institutions; establish a shared biorepository.	Promote use of implementation science methods to optimize utilization and scaling up of evidence-based, person-centered symptom interventions.	Establish a collaborative network and regular meetings; share a vision and mission with members of GRASS; and build the capacity to conduct symptom science research, education, and implementation projects on a large scale.

PRO—patient-reported outcome; WG—working group

Regarding GRASS’s organization and structure, founding members developed the following vision and mission statements:

• ■ Vision: to alleviate the global burden of chronic disease-related symptoms

• ■ Mission: Across the globe, throughout every population affected by any chronic health condition(s), we aim to:

  1. Facilitate a standardized approach to data collection to investigate symptoms.

  2. Understand symptom experiences, determine their underlying mechanisms, and identify effective, acceptable, and scalable treatment approaches.

  3. Facilitate implementation of evidence-based symptom management interventions.

  4. Serve as a platform for scientific exchanges and dissemination.

  5. Educate and mentor the next generations of symptom scientists in precision health and whole person research methods.

Based on the challenges, opportunities, and strategies identified by the WGs, five main task forces will address these five aims (see Figure 2).

FIGURE 2.

Vision, Mission, Aims, and Task Forces Based on the Identified Gaps and Strategies by the Working Groups

Taken together, we envision that GRASS will “grow” into a collaborative, inclusive, highly productive international consortium focused on generating new symptom science knowledge and translating those research findings into clinical practice across the world. With the seeds that this global alliance will sow, our current directions include to establish regular scientific meetings, foster interprofessional collaboration, and engage with symptom scientists across specialties. Embedded throughout is the aim to build greater capacity for symptom science through education, mentorship, and collaboration.

Funding Statement

This research was funded, in part, by a National Cancer Institute grant (CA286967; principal investigator: Harris).