Team Science Principles Enhance Cancer Care Delivery Quality Improvement: Interdisciplinary Implementation of Breast Cancer Screening Shared Decision Making

Abstract

PURPOSE:

Implementing shared decision making (SDM), recommended in screening mammography by national guidelines for women age 40-49 years, faces challenges that innovations in quality improvement and team science (TS) are poised to address. We aimed to improve the effectiveness, patient-centeredness, and efficiency of SDM in primary care for breast cancer screening.

METHODS:

Our interdisciplinary team included primary and specialty care, psychology, epidemiology, communication science, engineering, and stakeholders (patients and clinicians). Over a 6-year period, we executed two iterative cycles of plan-do-study-act (PDSA) to develop, revise, and implement a SDM tool using TS principles. Patient and physician surveys and retrospective analysis of tool performance informed our first PDSA cycle. Patient and physician surveys, toolkit use, and clinical outcomes in the second PDSA cycle supported SDM implementation. We gathered team member assessments on the importance of individual TS activities.

RESULTS:

Our first PDSA cycle successfully generated a SDM tool called Breast Cancer Risk Estimator, deemed valuable by 87% of patients surveyed. Our second PDSA cycle increased Breast Cancer Risk Estimator utilization, from 2,000 sessions in 2017 to 4,097 sessions in 2019 while maintaining early-stage breast cancer diagnoses. Although TS activities such as culture, trust, and communication needed to be sustained throughout the project, shared goals, research/data infrastructure support, and leadership were more important earlier in the project and persisted in the later stages of the project.

CONCLUSION:

Combining rigorous quality improvement and TS principles can support the complex, interdependent, and interdisciplinary activities necessary to improve cancer care delivery exemplified by our implementation of a breast cancer screening SDM tool.

INTRODUCTION

Breast cancer is a leading cause of death for women age 35-49 years.¹ Mammography reduces breast cancer mortality for women age between 40 and 49 years.^2-5 However, screening mammography is not without negative consequences such as increased health care utilization and patient anxiety because of false positives and benign biopsies.^6-9 Shared decision making (SDM) is recommended in national guidelines for women age 40-49 years.^2,4 The implementation of SDM for these women faces myriad challenges in practice,^10,11 revealing a familiar gap: cancer care delivery requires effective and high-performing teams to implement patient-centered programs.¹²

In this article, we highlight key quality improvement (QI) and team science (TS) principles used to help an interdisciplinary team implement voluntary SDM in primary care visits for breast cancer screening in women age 40-49 years. We used an iterative plan-do-study-act (PDSA)¹³ framework with a TS lens focused on shared goals, communication, culture, research support, data management, and leadership¹⁴ to codify and then implement an electronic health record (EHR)-embedded tool called the Breast Cancer Risk Estimator (B∼CARE). Our project had two aims: (1) to establish B∼CARE 1.0, a web accessible and EHR-embedded SDM tool that presents breast cancer risk factors and facilitates SDM for screening mammography in primary care focusing on the outcomes of patient/provider acceptability (patient satisfaction and provider satisfaction) and (2) to revise B∼CARE 1.0 and implement B∼CARE 2.0, focusing on utilization in clinical settings and diagnosis of early-stage breast cancer.

METHODS

Our QI project was conducted in the University of Wisconsin (UW) academic medical center composed of the health system (UW Health) and the medical school (UW School of Medicine and Public Health), which exists within the culture of the University (UW-Madison).

Team Science

At UW, our team was on the leading edge of developing an ecosystem in which clinicians and researchers (Data Supplement, online only) use PDSA¹³ cycles to iteratively improve quality.^15,16 Medically trained team members included a radiologist specializing in breast imaging and decision support (E.S.B.); a family medicine physician, specializing in women's health (S.S.); and a clinician who has integrated decision aids into EHRs (J.K.). Clinicians were joined by experts in population health science and breast cancer policy modeling (A.T.-D.), cancer communication and digital media (D.S.), simulation modeling and optimal health care policy (O.A.), and health psychology (L.D.). The team was complemented by strong and consistent project management¹⁷ (T.L.) to support communication, coordination tools, trainings, and process redesign.^18,19 We were supported by advocates and leaders across the enterprise, including patients²⁰ and clinicians.

During intense implementation periods, our team met every other week and communicated frequently by e-mail. We used strategies to convene small groups with a specific collection of the most relevant skill sets for each meeting, which added to efficiency—further attesting to the value of our project management lead (T.L.), who matched the meeting attendees to agendas and desired meeting outcomes. Coordination of expertise was the responsibility of the leads. However, each team member championed their unique area of expertise.

Our team had high task interdependence,²¹ and thus, at the start in 2016, we implemented TS best practices as soon as they became available in the literature.^12,21-26 For this summary of our QI efforts, we extracted six TS principles from the literature,¹⁴ documented to increase rigor and reproducibility, and collapsed them into four TS best practice dimensions, for simplicity. Specifically, from the study by Rolland et al,¹⁴ we left the first (mission, vision, and goals) and the sixth (leadership) TS best practices intact. We combined the second and third TS best practices (culture of trust and communication) and the third and fourth (operations management and data management) to enable our team to match these four TS best practice dimensions (Table 1) with our PDSA process over time.

TABLE 1.

Team Science Best Practices Associated With Project Activities at Each Stage of the PDSA Cycles According to Team Member Assessment/Consensus

Our team asynchronously discussed the mapping of these four TS best practice dimensions to the PDSA cycles and phases of each cycle. The lead author (E.S.B.) consolidated the group consensus into Table 1, which was then checked independently by each of the authors who fed back their comments. Any disagreements were discussed and resolved, thereby reaching consensus regarding which TS behaviors were most important for each PDSA cycle and phase.

Quality Improvement

PDSA cycle 1 (January 1, 2013-December 31, 2015): Our team collaboratively developed an application for pilot funding to support development of B∼CARE 1.0 for risk-based mammography SDM. We developed B∼CARE 1.0 in partnership with a patient advisory committee²⁷—see the Data Supplement for more details, interviews with 17 representative primary care providers (six internal medicine, eight family medicine, and three obstetrics/gynecology), and published literature related to communication techniques²⁸ and risk stratification methods²⁹ (Plan). Eleven providers were invited to test B∼CARE 1.0 (Do). After these providers engaged in conversations with 51 women considering mammography using B∼CARE 1.0, we surveyed these patients and providers to assess feasibility, satisfaction, and potential outcomes^29-31 (Study). On the basis of patient^30,32 and provider³⁰ input, we made modifications to B∼CARE 1.0 (Act) in advance of implementation (Fig 1).

FIG 1.

Graphical depiction of the activities in each of the two PDSA cycles. B∼CARE, Breast Cancer Risk Estimator; CME, Continuing Medical Education; MOC, Maintenance of Certification; PDSA, plan-do-study-act.

PDSA cycle 2 (January 1, 2016-December 31, 2018): Our team codified core components of breast cancer screening SDM via a Delphi³³—see the Data Supplement for details of patient/provider contributions, the development of a Toolkit,³⁴ and alignment with UW Health clinical SDM guidelines in preparation for widespread clinical use of B∼CARE 2.0 (Plan). During the B∼CARE 2.0 rollout, we aligned organizational leadership to support a dedicated educational campaign for all primary care providers (Do). We gathered ongoing acceptability, utilization, and clinical outcomes data including screening rates, early-stage (Surveillance, Epidemiology, and End Results definition: stage in situ and localized) cancer diagnosis, and provider surveys (Study). To integrate and communicate best practices and prepare for further continuous improvement, interprofessional Continuing Medical Education (CME) and Maintenance of Certification (MOC) modules were developed and made available (Act). We also measured B∼CARE toolkit downloads and CME/MOC participation.

RESULTS

Retrospective analysis demonstrated the feasibility of B∼CARE on our patient population.²⁹ Patient survey results indicate that 87% (45 of 51) found use of B∼CARE 2.0 to be valuable when discussing screening mammography with their primary care provider. Patient-identified challenges include the need for explanation of complex terms such as overdiagnosis and false positives. All primary care providers reported that B∼CARE 2.0 facilitated SDM, noting that graphic representations of personal risk/benefit assessment improved shared understanding (eg, of breast cancer risk and of possible screening outcomes such as true and false positives) and knowledge transfer. Provider-identified barriers included not enough time for discussion although having the SDM available in the EHR helped.^28,30,31 We also found that primary care provider framing of risk and potential outcomes during the SDM interaction influenced patient satisfaction.³¹

Team consensus (Table 1) revealed that TS principles varied by the PDSA cycle and phase. For example, we needed to establish and reinforce shared goals²² clearly and more intentionally in the first PDSA cycle, work that carried forward to the second cycle. Precise, timely, and accurate²¹ communication and intentional trust building were important for both PDSA cycles and all phases. Research management and data management, once established in the first PDSA cycle, were leveraged and required less effort in the second PDSA cycle. Our sustained, interdisciplinary leadership team, especially the co‐leaders from radiology and family medicine (E.S.B. and S.S., respectively), invested time in learning and implementing transformational leadership^35,36 techniques to support evidence-based TS best practice dimensions, which also extended to cycle two, with focused attention on plan and act.

Utilization of the B∼CARE tool, available in the EHR and online (Fig 2), increased steadily between 2016 and 2019 without major changes in the underlying clinical population. In 2017 compared with 2019, the two full years available for cumulative analysis, B∼CARE use more than doubled going from 2,000 to 4,097 sessions either through the EHR or online. The Toolkit³⁴ was downloaded 261 times by both local and international audiences. During the same time, health system mammography screening rates, for all medically homed patients, remained stable and the proportion of early-stage cancers remained steady or increased (Data Supplement). Fifty-three primary care providers accessed the CME module. Twenty-seven completed the module for credit (51%), all of whom proceeded to the MOC exercise.

FIG 2.

Monthly SDM sessions recorded after implementation of the B∼CARE tool in July 2016. B∼CARE, Breast Cancer Risk Estimator; SDM, shared decision making.

DISCUSSION

Relying on TS principles relevant to cancer prevention and control and established QI techniques, we developed and implemented SDM for breast cancer screening targeting 40- to 49-year-old women. We found that in the first PDSA iteration, we needed to rigorously apply the TS principles, which we could relax slightly in the second PDSA iteration except for continuous, intentional communication and trust building. One key to our success is that our tool, available in the EHR, is convenient for busy providers. Many similar tools are only web-based⁴¹ and thus not integrated in provider workflow in the EHR. Our investment in the creation of a toolkit and CME/MOC modules supports robust local dissemination, implementation, and sustainability as well as increasing the chances of generalizability and use in other clinical practices.

Several challenges and opportunities remain. In the future, we plan to better determine the utilization rates and effectiveness of SDM and more directly tie B∼CARE use to breast cancer outcomes. In addition, mammographic screening⁴² (and likely SDM rates, although this is beyond the scope of this article) decreased substantially during disruptions because of the COVID-19 pandemic, resulting in screening delays and likely future worsened breast cancer outcomes. Optimizing B∼CARE to support primary care providers' ability to deliver effective, patient-centered, timely SDM for breast cancer screening will be crucially important.

In concert with successful projects like ours, the UW academic medical center has increasingly adopted PDSA^43,44 to advance our learning health system approaches to deliver clinical and operational improvement and generalizable knowledge. The literature supports using TS advances in cancer care delivery including SDM implementation⁴⁵ and EHR interventions⁴⁶ to support high-performing teams to implement patient-centered programs.¹² Merging TS and PDSA QI cycles underpins the iterative Data to Knowledge, Knowledge to Practice, and Practice to Data⁴⁷ paradigm key to advancing learning health system scalability, catalyzing evidence-based⁴⁸ cancer care delivery, and achieving improved patient outcomes.

ACKNOWLEDGMENT

The authors acknowledge support from the University of Wisconsin Institute for Clinical and Translational Research supported by the Clinical and Translational Science Award (CTSA) program, through the NIH National Center for Advancing Translational Sciences (NCATS) grant (UL1TR000427), the University of Wisconsin Carbone Comprehensive Cancer Center Support grant (P30CA014520), the UW School of Medicine and Public Health Wisconsin Partnership Program, and the UW Health Innovations Program. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH or the other funding sources listed above.

The authors would also like to thank Lee G. Wilke, MD, FACS, Division of Surgical Oncology, Department of Surgery, University of Wisconsin for her contributions to the development of the Breast Cancer Risk Estimator (B∼CARE) and Toby Campbell, MD, MS, Division of Hematology/Oncology, Department of Medicine, University of Wisconsin for his contributions to the interprofessional Continuing Medical Education (CME) and Maintenance of Certification (MOC) modules. We would also like to thank the staff of the Wisconsin Research and Education Network for their support of the work conducted throughout the project and Robert Millholland for assistance with cancer registry data. We are extremely grateful for the time and expertise shared by these individuals.

AUTHOR CONTRIBUTIONS

Conception and design: Elizabeth S. Burnside, Sarina Schrager, Lori DuBenske, Amy Trentham-Dietz, Dhavan Shah, Oguzhan Alagoz

Administrative support: Elizabeth S. Burnside, Terry Little

Provision of study materials or patients: Sarina Schrager, Amy Trentham-Dietz

Collection and assembly of data: Elizabeth S. Burnside, Sarina Schrager, Lori DuBenske, Jon Keevil, Terry Little, Amy Trentham-Dietz, Dhavan Shah, Oguzhan Alagoz

Data analysis and interpretation: Elizabeth S. Burnside, Sarina Schrager, Jon Keevil, Amy Trentham-Dietz, Betsy Rolland, Dhavan Shah, Oguzhan Alagoz

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

Team Science Principles Enhance Cancer Care Delivery Quality Improvement: Interdisciplinary Implementation of Breast Cancer Screening Shared Decision Making

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/op/authors/author-center.

Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).