Advancing rapid cycle research in cancer care delivery: a National Cancer Institute workshop report

Abstract

Generating actionable research findings quickly and efficiently is critical for improving the delivery of cancer-related care and outcomes. To address this issue, the National Cancer Institute convened subject matter experts, researchers, clinicians, and patients for a 2-day virtual meeting in February 2022. The purpose of this meeting was to identify how rapid cycle interventional research methods can be used to generate findings useful in improving routine clinical practice. The meeting yielded an initial conceptualization of rapid cycle interventional research as being comprised of 6 key elements: use of iterative study designs; reliance on proximal primary outcomes; early and continued engagement with community and clinical partners; use of existing data sources to measure primary outcomes; facilitative features of the study setting and context; and consideration of appropriate rigor relative to intended use of findings. The meeting also identified the types of study designs that can be leveraged to conduct rapid cycle interventional research and provided examples of these; considered this approach from the perspective of key partners; described the clinical and data infrastructure, research resources, and key collaborations needed to support this work; identified research topics best addressed using this approach; and considered needed methodological advances. The National Cancer Institute is committed to exploring opportunities to encourage further development and application of this research approach as a means for better promoting improvements in the delivery of cancer-related care.

The last few years have witnessed an increase in interventional research funded by the National Cancer Institute (NCI) and other organizations focused on improving the delivery of care across the cancer control continuum (1). This increase coincides with the growth of health-care delivery research, a multidisciplinary effort which recognizes that care delivery occurs in a multilevel environment encompassing patients, practitioners, health-care systems, and communities. Recent examples include projects funded under the NCI Cancer Moonshot designed to improve rates of colorectal cancer screening among people who are medically underserved (ie, Accelerating Colorectal Cancer Screening and follow-up through Implementation Science; https://healthcaredelivery.cancer.gov/accsis/]) and to improve symptom assessment and management in people diagnosed with cancer (ie, Improving the Management of Symptoms during And following Cancer Treatment; https://healthcaredelivery.cancer.gov/impact/]).

Despite the increase in research activity, observers have criticized the pace of progress in improving the delivery of cancer care (2), which may take 7 years or more to move from concept development to generation of findings applicable to clinical settings. Implementing these interventions can take even longer for populations that are underserved by health-care services and settings that are underresourced (3). A similar criticism has been offered for the relatively slow pace of developing (or adapting) and testing interventions that could improve cancer-related outcomes (4). These critiques point to the need to increase the pace and efficiency of generating findings that can be used to improve delivery of cancer-related care and outcomes.

Several recent advances suggest ways these criticisms can be addressed by applying a rapid cycle methodological approach to interventional research. One advance is the delineation of a learning health system model in which internal, routinely collected data and experiences are systematically integrated with external evidence, with that knowledge then put into practice (5). A related development is progress in the aggregation and use of clinical data from the electronic health record (EHR) for research and quality improvement purposes (6). Advances in conceptualizing and designing randomized controlled trials, most notably pragmatic trials, are being used more often as a way to improve efficiency, accelerate progress in answering research questions, and generating evidence that more accurately reflects the settings to which the results would apply. Indeed, tools have been developed to facilitate structured planning for pragmatic trials, including the Pragmatic Explanatory Continuum Indicator Summary-2 for individual- or patient-level trials (7) and the Pragmatic Explanatory Continuum Indicator Summary-2-Provider Strategies for provider- or organizational-level trials (8). Lastly, there is an increasing emphasis and recognition of leveraging engagement research to systematically and collaboratively involve key partners, such as patients, health-care providers, and health-care system leaders in all aspects of research, from identification of priority research questions through generation and dissemination of study results (9).

To generate greater interest and awareness of the potential for rapid cycle interventional research to advance cancer care delivery research, the NCI Division of Cancer Control and Population Sciences sponsored a virtual workshop, “Advancing Rapid Cycle Research to Improve Cancer-Related Care,” on February 16^th and 17^th, 2022. Toward this end, the meeting brought together researchers, practitioners, policy makers, health-care system leaders, and patients. The focus was on how rapid cycle interventional research could be leveraged to accelerate the generation of evidence by answering multiple research questions within the context of a single study, rather than the typical approach of answering a single research question within a single study.

Origins of rapid cycle research

It is only fairly recently that “rapid” has been called out as a useful approach to conducting health-care research. Among the earliest articles to focus on the need for rapid or rapid cycle interventional research in health-care settings are 3 publications from 2012, 2013, and 2014, respectively.

The first article (10) makes the point that scientific evidence has generally not translated rapidly or consistently into policy and practice. To address this issue, the authors call for and make recommendations about use of research methods that are rapid, rigorous, transparent, and contextually relevant. Four issues are identified that, if addressed, could contribute to making health research more applicable to health-care settings in a timely manner: 1) studying issues in context; 2) a more pragmatic or realist orientation to research; 3) an approach to study design that balances considerations of rigor with relevance; and 4) a flexible approach toward investigation that is adaptive, iterative, and evolving.

The second article (11) makes a similar point that research often fails to find its way into practice or policy in a timely way, if at all. The authors offer several recommendations to address this issue, including making research more relevant to community and clinical collaborators; using approaches that are rapid and recursive, such as allowing discoveries within a study to influence the study; redefining rigor as the property of a series of decisions, observations, and relationships rather than a list of techniques; reporting resources used to conduct the study or replicate the interventions including costs; and designing for replicability and reporting features of the study that would be needed to replicate it under the same or different conditions.

The third article (12) reiterates the point that progress in health-care research is slow and results seldom translate into practice. The authors directly address the importance of moving to a rapid research model and identify 4 considerations as important for moving to that model. The first issue is stakeholder engagement, which they acknowledge may seem counterintuitive as a strategy to accelerate research. Instead, they view this approach as having the potential to improve participant recruitment and retention, as well as making the research process more efficient, thereby increasing the likelihood that findings will be relevant to partners and more readily translated into practice. A second issue is greater adoption of approaches characterized as rapid research designs. This includes sequential multiple assignment randomized trial [SMART (13)] designs to test multiple intervention components and factorial designs for identifying and fine-tuning components of effective interventions. A third issue is a call for more rapid review processes that streamline grant review and funding activities, and place a greater premium on use of more rapid and innovative research designs. Lastly, the authors describe aspects of research infrastructure that can facilitate rapid research, such as leveraging the EHR as a data source, and making greater use of common data elements to improve efficiency and facilitate data sharing.

Subsequent to these publications, the Agency for Healthcare Research and Quality released a report in 2015 titled, Using Rapid-cycle Research to Reach Goals: Awareness, Assessment, Adaptation, Acceleration (14). The authors summarize how methods for pursuing rapid cycle research are emerging as quality improvement teams address clinical problems, research teams study implementation, and recognition grows of the need for prompt assessment and actional adaptations. The report describes a 6-phase model for conducting rapid cycle research comprised of preparation, problem exploration, knowledge exploration, solution development, solution testing, and implementation and dissemination.

Taken together, this relatively recent body of work suggests there is growing interest in leveraging rapid cycle interventional research to improve health-care delivery. It also illustrates how this type of research is informed by and relevant to several scientific disciplines, including health-care delivery research, implementation science, and quality improvement research.

Overview of workshop goals

The overall objective of the NCI workshop was to increase awareness and interest in conducting rapid cycle interventional research and its potential to optimize cancer care delivery. Specific meeting goals were to 1) solicit feedback on proposed key elements of rapid cycle interventional research in cancer care delivery; 2) identify the types of study designs that can be leveraged to conduct this research; 3) evaluate rapid cycle interventional research from the perspective of key community and clinical stakeholders; 4) identify the clinical and data infrastructure, research resources, and key collaborations needed to conduct rapid cycle interventional research; 5) identify pressing and understudied research questions best answered using such designs; and 6) identify methodological applications and advances needed to support this work.

Several formats were used to achieve meeting objectives, including panel presentations and small breakout group discussions and report backs, interspersed with question and answer segments. An archive of the workshop agenda, workshop report, session recordings, summary of breakout group discussions, and other materials is available online (https://healthcaredelivery.cancer.gov/rapid-cycle/).

Key characteristics of rapid cycle interventional research

The NCI workshop began with a presentation of an initial conceptualization of key characteristics of rapid cycle interventional research. In offering this conceptualization, several points were noted. First, the proposed elements represented initial thinking among the workshop organizers, and one of the goals of the workshop was to refine this list. Second, we recognized that the individual characteristics described are not unique to rapid cycle interventional research. Rather, it is the combined presence of several of these elements that results in research being consistent with this approach. Finally, the characteristics are not intended to be viewed as a checklist to determine if the interventional study qualifies as rapid cycle research. Rather, evaluation of the presence of these elements can be used to help identify the extent to which an interventional study possesses these characteristics.

Identification of these key characteristics was based on review of recent publications and reports that have advocated for the adoption of more rapid approaches to conducting health-care delivery research (10-12,14). It also reflects discussions among planning committee members, including a working definition of rapid cycle interventional research: a rigorous approach to conducting interventional research that seeks ways to maximize the timeliness and efficiency of the process for generating answers to questions of practical interest and actionable use. With these considerations in mind, the initial characterization of rapid cycle interventional research consists of 6 key elements.

The first and perhaps most unique characteristic of rapid cycle interventional research is the use of iterative study design features. This element can be reflected in plans to complete at least 2 consecutive iterations or 2 concurrent comparisons of interventions. Consecutive iterations are a design in which results from the first cycle inform how to proceed in the second cycle (either a priori or post-hoc decision), with a similar strategy followed if there are subsequent cycles. For example, this approach could be applied to examine the best form of care for individuals who are nonresponders to an initial intervention aimed at improving adherence to recommended cancer screenings, or used to examine successive efforts of a health-care system to improve rates of completing recommended cancer screenings in their patient population. Concurrent comparisons are a full factorial design in which a study tests the effects of all combinations of the presence or absence of different intervention components to identify which components are making positive contributions to the overall intervention effect. For example, this approach could be used to identify the optimal combination of educational, psychological, and pharmacological strategies to include in a multicomponent intervention to promote tobacco cessation among current smokers recently diagnosed with cancer. In these examples, no time frame is specified for what qualifies as rapid; rather, these approaches have the potential to more efficiently and, therefore, more rapidly answer multiple scientific questions in a single study rather than answering a single scientific question in a single study or multiple questions answered in multiple studies.

The second characteristic of rapid cycle interventional research is a focus on proximal primary outcomes, as in outcomes that would be expected to be influenced in the shorter term by an intervention. This approach has the potential to yield earlier answers to questions of interest or an earlier read on the likelihood of the intervention having the desired effect. For example, in examining the impact of an implementation strategy seeking to promote greater adoption of evidence-based cancer treatment guidelines, one might focus on a health-care process indicator (eg, verified delivery of the evidence-based care) rather than a patient outcome indicator (eg, patients’ disease status years later). In many instances, selection of a proximal primary outcome will be based on the extent to which there is evidence that it is strongly correlated with or predictive of more distal outcomes of importance. In this example, one might select among possible proximal outcomes based on the strength of the evidence that a particular aspect of guideline-concordant care delivery leads to better health outcomes. In addition to process indicators, several outcome indicators can be identified that may be suitable as proximal primary outcomes. Depending on the research question, these may include short-term clinical events (eg, rates of post-discharge rehospitalization) or patient-reported health states (eg, ratings of nausea severity during chemotherapy administration).

A third element is an emphasis on partner engagement to conduct research and generate results that are relevant and applicable to cancer care delivery (15). For example, in designing a system to collect symptom reports from patients receiving chemotherapy to improve symptom control, early engagement of patients, clinical staff, health information technology (HIT) staff, and others has the potential to reduce the need for time-consuming redesigns, and yield a system more likely to be adopted. From a rapid cycle perspective, engagement can also have a positive impact on identification of priority topics and selection of an overall study design, study outcomes, and initial and subsequent interventions. All this suggests there are multiple ways by which engaging partners in the research process can increase the efficiency and reduce the time required to complete a study and apply its findings.

A fourth element is a consideration of data sources. From a rapid cycle interventional perspective, the focus is on the potential to use existing data as the source for measuring a primary proximal outcome rather than engaging in original data collection. Accordingly, investigators should explore whether valid and reliable data are available from existing sources (eg, EHR, administrative records, or clinical operations). For example, a study seeking to improve management of chemotherapy-related mucositis might rely on pharmacy data for information on administration of guideline-concordant topical agents and clinical notes for information on the occurrence and severity of patients’ mucositis to measure the outcomes of interest.

A fifth element is consideration of the setting in which the study will be conducted. In this regard, it is important to specify the infrastructure needed to support rapid cycle interventional research. Components include the clinical and research expertise needed to conduct this type of research and organizational capabilities for collecting and using data that can improve clinical practice. Regional oncology collaboratives focused on improving cancer care, such as the Michigan Oncology Quality Consortium (16), represent promising settings for conducting rapid cycle interventional research in cancer care delivery. In describing facilitative infrastructure, it is critical to identify ways to conduct this type of research in settings that may require additional resources. It is also important to identify designs and methodological approaches that can minimize the resources needed to answer questions of interest.

The sixth and final element of rapid cycle interventional research is consideration of appropriate rigor. In seeking ways to generate clinically useful results in a timelier manner, investigators should balance the rigor of the design best suited to answering the research question against considerations of the methods that are feasible in the study setting and the purpose for which the results will be used. Investigators should evaluate the pros and cons when choosing among available rapid cycle interventional designs (17). This includes weighing the potential for bias and threats to internal validity against the generalizability and applicability of study findings. For example, Vaisson and colleagues (18) used a rigorous randomized design (ie, 2x2x2 factorial experiment) that was adequately powered (94%) to detect statistically significant differences between the effect of different messages on access to an audit and feedback tool and cancer screening rates among a sample of more than 5000 primary care physicians.

Following initial presentations of these elements, workshop participants provided feedback about the importance of these elements via a real-time survey. The feedback was generally positive. In addition, comments provided by discussants focused on 3 suggested modifications to the original proposed characteristics. First, consideration of settings should be broadened to contexts to reflect the circumstances and environment in which the research will be conducted. Second, greater emphasis should be placed on efficiency by using methodology that can optimize resources needed to answer research questions. Third, health equity should be viewed as a cross-cutting theme. Investigators should consider how to incorporate health equity within all 6 characteristics of rapid cycle interventional research. For example, investigators should consider including low-resourced settings in rapid cycle interventional research to maximize health benefits for all. Inclusion of equity-focused proximal primary outcomes that are selected by stakeholders as part of the partner engagement process would contribute to maximizing equitable impact. A revised version of the 6 characteristics incorporating these suggestions, and highlighting health equity as a cross-cutting theme, can be found in Figure 1.

Figure 1.

The 6 key characteristics of rapid cycle interventional research. The characteristics are 1) use of iterative study design features; 2) focus on proximal primary outcomes; 3) emphasis on partner engagement; 4) consideration of data sources; 5) consideration of the context in which the study will be conducted; and 6) consideration of appropriate rigor. Health equity is a cross-cutting theme that should be incorporated into all 6 characteristics.

Study designs for rapid cycle interventional research

Three presentations from expert methodologists provided a high-level overview of state-of-the-art study designs suitable for rapid cycle interventional research. The first presentation described the SMART design and identified how it can be used to study small, incremental changes to implementation strategies as they are adapted over time in response to the changing context of implementation. The second presentation provided several examples of the use of SMART designs in health-care delivery research. Examples included studies aimed at identifying optimal strategies for promoting adoption of evidence-based practices and programs, such as testing iterations of coaching and facilitation strategies to increase adoption of cognitive behavioral therapy in statewide school systems (19,20). These presentations discussed the benefits of using SMART designs, including timeliness of generating answers and specificity of research questions. Some challenges when using SMART designs were noted, including the number of sites needed to answer research questions and the need for specialized methodological expertise.

The third presentation provided an overview of quality improvement study designs, such as the Model for Improvement using plan-do-study-act cycles (21). The presentation included an example of how planned experimentation using factorial designs can be incorporated into plan-do-study-act cycles to identify prevention practices that contribute to reduced hospital infection rates (22). The presentation also touched on the benefits of using factorial designs, including the efficiency with which one is able to identify the optimal sequence and combination of components for specific outcomes, target populations, and contexts (23). A drawback of factorial designs includes the number of sites needed for iterative testing. Audience questions and subsequent discussion focused on the role of partners in selecting a study design, data requirements for conducting efficient and rigorous rapid cycle interventional research, mismatch in timing between grant funding and these types of studies, and suggestions for working with institutional review boards (IRBs) to minimize any delay in the research process.

Frontline reports from researchers

Next, a panel of researchers spoke about their experience conducting studies using rapid cycle interventional research approaches in health-care delivery settings. The examples cut across a variety of areas, including cervical cancer screening (18), flu vaccination (24), depression screening and referral (25), patient navigation (26), and pathological staging for lung cancer (27). Panelists shared lessons learned for conducting research in this area. One panelist encouraged engaging with IRBs during the research study development phase. Researchers may need to educate IRBs on rapid cycle interventional research, explain why a more flexible review and approval process is necessary for supporting this type of research, and provide examples of how this can be accomplished while maintaining ethical standards and oversight. Another panelist recommended budgeting for dedicated time for a health information technologist, ideally someone who is embedded within the health-care system rather than part of the research team exclusively. Because leveraging these complex data for research purposes can be challenging, including a research team member who is adept at navigating the EHR system was viewed as essential.

Stakeholder perspectives

A panel of invited stakeholders provided their views of rapid cycle interventional research from a range of perspectives, including patient advocacy, professional organization leadership, nursing, community-based cancer care, clinical informatics, and cancer care delivery research. In their remarks, they emphasized the importance of considering each of these perspectives in designing and conducting rapid cycle interventional research and in disseminating and applying the findings. Panel members also stressed inviting meaningful partner engagement and underscored the importance of adequately recognizing partner contributions.

Breakout groups

Toward the conclusion of the workshop, attendees participated in 1 of 4 breakout groups. During this time, participants were asked to brainstorm and discuss initial ideas related to rapid cycle interventional research in 1 of 4 topics: 1) clinical and data infrastructure, 2) key collaborations, 3) research resources, and 4) research questions and methodological considerations. Participants were asked to incorporate a health equity lens into discussions to maximize equitable impact.

Clinical and data infrastructure

Participants in this breakout group were asked to discuss clinical and data infrastructure characteristics that would best support the conduct of rapid cycle interventional research in cancer care delivery settings. Key aspects for data infrastructure included the ability to share data and study results with clinical partners; common data dictionaries; open-source algorithms and software; training; and local HIT experts. Resources for working with IRBs; creating a business plan for cancer care health systems and settings to conduct rapid cycle interventional research; support for navigating regulatory roadblocks; and dedicated time for clinical staff and system leaders to engage in rapid cycle interventional research were also noted.

Key collaborations

Participants in this breakout group were asked to reflect on key collaborations for rapid cycle interventional research and describe why and during what stage collaborators should be engaged. Collaborators included patients and family members, advocacy groups, HIT staff, health-care providers, community thought leaders, professional societies, and executive clinical care leadership. For example, patients and providers should be involved during the study design phase to select meaningful outcomes. A supportive culture for research-practice partnerships, characterized by high levels of psychological safety and meaningful incorporation of others’ ideas, was also noted.

Research resources

Participants in this breakout group were asked to identify resources needed to support rapid cycle interventional research in cancer care delivery settings. These included training for how best to communicate rapid cycle interventional research to IRBs to minimize lag between sequential cycles of intervention testing. Methodological training, toolkits, examples of study protocols, and textbooks were identified as resources needed to support this research. Participants also identified the need for training in how to effectively communicate study findings to collaborators and the broader practice community, as well as the need to secure monetary support for generating pilot data and ensuring long-term sustainability.

Research questions and methodological considerations

Participants in this breakout group were asked to brainstorm topic areas for rapid cycle interventional designs. Examples generated include effective models for patients’ transition from cancer care to primary care, patients’ engagement and retention in cancer care, and tailored communication messages to increase screening rates. Suggestions for advancing methodology included training institutes, repository of tools, textbooks, and consultation models. Guidance to help researchers decide when one type of rapid cycle interventional study design would be better than another was encouraged. Attendees emphasized the need to learn from natural experiments, as well.

Discussion

Further development and use of rapid cycle interventional research can help accelerate the timeliness and improve the efficiency of research aimed at optimizing delivery of cancer-related care and cancer outcomes. NCI convened a workshop that brought together individuals with relevant experience and interests to consider how to advance use of this approach. The workshop yielded a refined conceptualization of the key characteristics of rapid cycle interventional research to better define the area, identified study designs that can be leveraged to conduct this research, and provided examples of those designs. Key collaborators were identified, along with specific areas and phases of research where their involvement would be especially helpful. The workshop also characterized the clinical and data infrastructure and resources that can facilitate conduct of rapid cycle interventional research and identified several priority research topics. The NCI expects to use the insights from this workshop to develop a strategic plan for increasing the use of rapid cycle interventional research to improve cancer care delivery.

Funding

Not applicable.

Notes

Role of the funder: Not applicable.

Disclosures: The authors have no conflicts of interest to disclose.

Author contributions: Writing—original draft: PJ, WN; writing—review and editing: PJ, WN, AK, BM, GP, SS, RV, ET.

Acknowledgements: The authors thank the workshop planning committee, presenters, and attendees for their active participation, thoughtful comments, and contribution to the workshop. The authors also thank Trish Silber, Aliniad Consulting Partners, Inc, and staff from ICF Next for their invaluable workshop planning and support services.

Disclaimers: The observations and conclusions in this commentary are those of the authors and do not represent the official position of the National Cancer Institute, National Institutes of Health, or other US federal agencies.

Data availability

No new data were generated or analyzed in support of this manuscript.