Multi-pilot implementation experiences of patient-centered pathology reports: lessons learned for the advancement of patient-centered tools for cancer decision-making

New federal legislation in the United States grants patients expanded access to their medical records, making it critical that medical records information is understandable to patients. Provision of informational summaries significantly increase patient perceptions of patient-centered care and reduce feelings of uncertainty, yet their use for cancer pathology is limited. Our team developed and piloted patient-centered versions of pathology reports (PCPRs) for four cancer organ sites: prostate, bladder, breast, and colorectal polyp. The objective of this analysis was to identify common barriers and facilitators to support dissemination of PCPRs in care delivery settings. We analyzed quantitative and qualitative data from pilot PCPR implementations, guided by the RE-AIM framework to explore constructs of reach, effectiveness, adoption, implementation, and maintenance. We present two case studies of PCPR implementation – breast cancer and colorectal polyps—that showcase diverse workflows for pathology reporting. Cross-pilot learnings emphasize the potential for PCPRs to improve patient satisfaction, knowledge, quality of shared decision-making activities, yet several barriers to dissemination exist. While there is promise in expanding patient-centered cancer communication tools, more work is needed to expand the technological capacity for PCPRs and connect PCPRs to opportunities to reduce costs, improve quality, and reduce waste in care delivery systems.


Background
Cancer care is complex and fraught with myriad of treatment options that can be difficult to navigate. Given the complexity of cancer decision-making, national oncology organizations recommend routine use of shared decision-making (SDM) as a best practice and central element of high-quality cancer care [1][2][3][4]. SDM is predicated on patients having access to information that helps them understand their diagnoses and participate in treatment decision-making. Multiple studies have highlighted that SDM requires cultural and structural changes within healthcare delivery systems and that efforts to effectively implement SDM in practice are variable and evolving [1,2,[5][6][7]. One key challenge to effective implementation of SDM in cancer care is the complexity of cancer pathology reports that dictate treatment decisionmaking. A recent systematic review of barriers to SDM in cancer care identified that one of the most frequent barriers patients face is uncertainty in their treatment decision [1]. This can also serve as a barrier for healthcare professionals, who may struggle to engage patients in the SDM process when the patients feel less informed about their diagnosis [5].
The 21st Century Cures Act federally mandates that patients have immediate access to their medical records, including cancer pathology reports. It is therefore critical that we ensure pathology reports are understandable to patients or supplemented with patient-friendly informational summaries [8]. Provision of informational summaries following cancer care have significantly increased patient perceptions of patient-centered care and reduced feelings of uncertainty [9] during the important transition from the treatment to survivorship phases of care. Yet informational summaries at the initiation of cancer care are rare. Addressing this need will require the redesign and restructuring of electronic health record (EHR) tools and care delivery systems to be more inclusive of and oriented towards patient engagement [10]. Applying the principles of dissemination and implementation science may aid these complex practice changes and enhance our ability to provide more patientcentered care [11].
Over the past seven years, our team has developed and piloted patient-centered versions of pathology reports for four cancer organ sites: prostate, bladder, breast, and colorectal polyp [12][13][14][15]. Here, we reflect on cross-pilot learnings from these four pilot trials related to barriers and facilitators of patient-centered pathology report (PCPR) implementation. Using the example of PCPRs, our aim is to identify critical considerations for dissemination and implementation of communication tools and practices that can enhance patient-centered cancer care.

Pilot studies and data sources
Details of the pilot trials have been described previously and are summarized in Fig. 1 [12][13][14][15]. In brief, PCPR templates were iteratively developed for each organ site via multidisciplinary input from clinical experts (oncology, surgery, pathology), researchers, and patient advisors. The team then led four separate pilot randomized controlled trials to evaluate the impact of PCPRs on patient knowledge of their pathology and experiences with cancer decisionmaking (primary results reported elsewhere for each trial). Each pilot trial recruited patients with new cancer pathology and patients completed semi-structured surveys at two timepoints: initial receipt of pathology reports (standard pathology only for the control group, standard pathology and PCPR for the intervention group), and 30 days later. Patients received their pathology reports (standard and PCPR) electronically through a study-related portal; standard pathology reports were also available in the patient portal as part of usual care. Pilot trials occurred for PCPRs developed for bladder cancer (2016), prostate cancer (2019), breast cancer (2020) and colorectal polyp (2021). A total of 228 patients were recruited across all four pilot studies. Recruited patients had no prior cancer diagnosis, were English-speaking, and over 18 years of age.
The objective of this analysis was to leverage lessons learned from all four pilot trials to identify common barriers and facilitators that can support future efforts to disseminate  PCPRs in care delivery settings. Our analysis was guided by the RE-AIM framework, an established implementation science framework used to guide evaluation efforts at any stage of implementation [16]. RE-AIM considers the interrelated constructs of reach, effectiveness, adoption, implementation, and maintenance, each of which can inform efforts to disseminate evidence-based practices with potential to improve cancer care delivery [17]. The present analysis aggregated quantitative and qualitative data (Supplemental Table A) across the four pilot studies, including additional data collected but not reported in the initial studies, to identify metainferences for the scale and spread of PCPRs. Data sources included patient and healthcare professional-reported semistructured surveys gathered throughout pilot studies, data generated from PCPR testing and quality assurance activities, and fieldnotes from implementation team meetings across the pilots. Quantitative data were analyzed using descriptive statistics.
Qualitative data from open-ended survey responses were analyzed using directed content analysis. Fieldnotes from implementation team meetings were analyzed using a deductive thematic analysis guided by RE-AIM, with iterative review by full study team to confirm themes. The team reviewed quantitative and qualitative data within each pilot, and then across all four pilots, to identify cross-pilot learnings as well as learnings unique to each cancer site. Table 1 provides an operational definition of the RE-AIM constructs for PCPRs, as well as a summary of RE-AIM findings and recommendations. First, we present two case studies of PCPR implementation experiences. These two case studies-breast cancer and colorectal polyp-were chosen because they represent diverse workflows for pathology reporting in practice. A patient with breast cancer, for example, can receive pathology at several timepoints (e.g., core needle biopsy, postsurgical), and may incorporate multiple complex clinical concepts into their decision-making about cancer treatment. In contrast, a patient receiving a routine colonoscopy will use the results of their pathology to determine the frequency of their next screening colonoscopy and may or may not discuss these results with a healthcare professional directly. Following, we summarize how cross-pilot learnings can inform future dissemination for PCPRs and other patientcentered tools for cancer care.

Breast cancer pathology
Breast cancer PCPR templates were developed for core needle biopsy and surgical pathology reports. PCPR templates required eight rounds of quality assurance testing. For each round, a representative pathologist and surgical oncologist compared PCPR language with standard pathology report information and provided blind ratings of clinical accuracy for each cancer domain (e.g., stage, grade). Ratings were compared and arbitrated by the multidisciplinary team until consensus was reached on needed PCPR adaptations. PCPR templates for pathology of core needle biopsies reached consensus on clinical accuracy more quickly. In contrast, surgical PCPRs were more time intensive to refine as they included more complex clinical features (e.g., bilaterality, rare histology). Midway through PCPR development and testing, American Joint Committee on Cancer (AJCC) guidelines for breast cancer staging changed, necessitating a revision of PCPR language to reflect updated guidance.
PCPRs templates were built into PowerPath (Sunquest Information Systems, Tucson, AZ), a pathology laboratory information software (LIS) used by our healthcare system at the time of the pilot. PCPR templates followed the existing clinical templates for pathology reporting, so that at the conclusion of completing the standard pathology report, pathologists were given the option to generate an accompanying PCPR with a single click. Prior to pilot launch, the study was presented to all breast cancer pathologists onsite; however, not all pathologists elected to participate in the trial, citing beliefs that pathology reports are only intended for healthcare professionals, not patients. The exclusion of some pathologists from the pilot trial limited our reach of potential patients, particularly in recruitment for core needle biopsy patients, which has a rapid turnaround for pathology results (often 1-3 days).
During the pilot, 66 patients with a new breast cancer diagnosis (defined for this study to include invasive carcinomas or ductal carcinoma in situ) were enrolled, randomized to receive standard pathology reports with or without a PCPR, and completed study activities. Over the course of the study, PCPRs for 3 patients with surgical pathology (representing 8% of the PCPRs needed) could not be generated due to clinical complexity, including bilateral cancers and/ or challenges with how some clinical features (e.g., hormone receptor status) were reported in non-standard ways that could not be integrated into PCPR algorithms. Enrolled patients received their pathology reports prior to scheduled visits with their surgical oncology healthcare professional. In surveys (n = 56), surgical oncologists indicated that almost 90% of patients did not utilize a copy of their pathology reports (either on paper or digital) during their appointment. However, surgical oncologists reported that a higher proportion of patients in the PCPR arm referred specifically to their pathology results compared with patients who only received the standard pathology report (60% vs. 47%). Surgical oncologists also reported that they used the PCPR during the visit to help clarify terminology and concepts approximately 20% of the time. Within the survey, surgical oncologists offered comments that "the PCPR seemed to make the patient more prepared to discuss the path report" and that patients who received a PCPR "came in more educated with more directed questions [regarding] next steps of therapy as determined by the path report." One surgical oncologist described an example, saying: "The patient and her husband found [the PCPR] extraordinarily helpful. They expressed that the conventional pathology report contained a lot of medical jargon that they didn't understand, but that the PCPR clarified" Among patients in the study, overall knowledge was higher in the PCPR arm compared with the control arm (66% vs. 50%); however, knowledge gains decreased over time and differences did not reach statistical significance. At the 30-day survey timepoint, 95% of patients in the PCPR arm agreed or strongly agreed that the PCPR was easy to understand, 90% reported that the PCPR helped them better understand their results, and 79% believed that it helped them better communicate with their healthcare professional. Open-ended responses (n = 41) on patient surveys indicated that patients felt the PCPRs "expanded understanding" of their pathology. One patient said "I liked that the report explained what different terminology meant and explained what various scores meant. Very helpful." Another said, "I really like the patient centered pathology report and will bring it with me when meeting with my oncologist." In contrast, patients in the control group who only received standard pathology reports cited the need to have "a decoder for the different technical terms" and "more clear explanation on some technical terms." As one patient who only received the standard pathology report described: "The version I received in [the patient portal] was all crammed together with odd line breaks, making it hard to read. A key to findings would be helpful, e.g., listing 'Not performed,' 'Not identified,' etc. along with an key as to whether they mean the same thing, and the implications of those items as they relate to terms like 'positive' or 'negative.'" Another patient who only received the standard pathology report echoed this frustration, saying: "There should be a glossary written in plain language. For example, the Nottingham score? I googled everything AND consulted with a physician friend of mine.
[…] A glossary is needed along with context. If something is grade 1, how many grades are there? […] It's easier for me to see and digest the info THEN talk it over with my doctor."

Colorectal polyp pathology
PCPRs for colorectal polyps were developed for noncancerous, precancerous, and cancerous polyp pathology, which reflected the breadth of findings from screening colonoscopies. The team met preliminary barriers, as polyp pathology often involves pathology results for multiple individual specimens, making it difficult to generate collective findings from multiple specimens in an automated way. Midway through PCPR development, our health system transitioned to a new EHR system, and the team was unable to prioritize health information technology (HIT) resources to build PCPRs within the new laboratory information system. Therefore, the team utilized a template outside of the EHR to generate PCPRs in a semi-automated fashion during the pilot study.
During the pilot, 44 patients were enrolled, randomized to receive pathology reports with or without a PCPR, and completed study activities. Patients received their pathology 1-2 weeks following their colonoscopy and typically did not have a follow-up visit with their healthcare professional. Overall knowledge of polyp pathology was slightly lower for the PCPR group at baseline (70% vs. 76%). However, knowledge was much higher (66% vs. 41%) for the PCPR group at the 30-day retention timepoint. The two domains patients with PCPRs had the highest knowledge for were understanding the types of polyps they had and the timing of their next screening colonoscopy.
At the 30-day survey timepoint, 85% of patients in the PCPR arm agreed or strongly agreed that the PCPR was easy to understand, and 70% reported that the PCPR helped them better understand their results. Open-ended responses (n = 28) from patients emphasized that the PCPR "had nice, simple, and relevant language with outcomes clearly spelled out […] and I didn't feel like I needed to do research to understand what it meant." As one patient described, "communication to patients should be optimized in a separate report; so I would support the effort to have a separate report for patients." Again, patients in the control arm who only received the standard pathology report highlighted that pathology reporting "need[s] to be written for a non-medical person." As one patient said, "it is frustrating that I have to do so much internet research to find out what the diagnosis means."

Dissemination and implementation learnings
The goal of each pilot trial was to assess the efficacy of PCPRs for specific cancer sites; in this analysis, we used implementation data from all pilots to assess the feasibility 1 3 of scaling PCPRs across organ sites, pathology types, and care settings. Cross-pilot learnings are described in the following sections. Table 1 provides RE-AIM operational definitions and a summary of learnings and recommendations for the expansion of patient-centered tools for pathology reporting.

Reach
Through the pilots, we developed PCPR templates for four of the most common cancer sites: prostate, bladder, breast and colorectal polyp. The range in organ site and pathology reporting type across these four pilot trials demonstrate the potential for PCPRs to be utilized across the continuum of cancer treatment decision-making. Yet we also found that the ability to automate PCPRs decreased as clinical complexity increased. In the example of breast cancer, we were unable to generate PCPRs for some patients with surgical pathology during the pilot trial due to the complexity of their pathology features; however, this barrier was not met for core biopsy pathology. Overall, there is promise in extending the reach of PCPRs to all organ sites, but more work is needed to ensure PCPRs are able to reach all patients.

Adoption
Learnings from these pilot trials foreshadowed potential adoption barriers when PCPRs are used in real-world settings. For example, some pathologists had negative beliefs about using PCPRs, even in the context of a pilot trial. Additionally, few patients incorporated their pathology reports into clinical appointments. These findings align with other work [9] that found that patients often do not access summaries intended to improve the patient-centeredness of care delivery, even when offered, and emphasize the need for more targeted stakeholder engagement work to address anticipated barriers to adoption. Additionally, future work is needed to explore the perspectives of other healthcare professionals beyond surgical oncologists related to feasibility and adoption of PCPRs in routine practice.

Effectiveness
We considered demonstration of effectiveness of PCPRs across pilots to be an increase in patient knowledge for their pathology (proximal outcome) and enhancing activities of shared decision-making with their clinical teams (distal outcome). Across all four pilots, PCPRs demonstrated increases in patient knowledge for their pathology compared with standard pathology reports. The majority of patients cited that PCPRs enhanced their ability to understand their diagnosis and communicate with their clinical team. Qualitative feedback from patient surveys confirmed their appreciation for the simple language, clear formatting, and ability to share with family and caregivers. Those that only received the standard pathology report reiterated the challenges identified in prior formative work around the need for pathology reports that provide plain language summaries and patientfriendly formatting [18]. However, given the small sample sizes of our pilots, we were unable to meaningfully explore potential differences by patient demographic characteristics. Future studies should expand to more diverse settings and patient populations to evaluate the comparative effectiveness of PCPRs at different timepoints in a new cancer diagnosis (i.e., initial pathology vs. definitive surgical pathology), as well as across demographic subgroups.

Implementation
We found that replicating PCPR template language and learnings from multiple pilots increased efficiency of PCPR development across organ sites. Over the course of the pilot trials, though, we encountered several barriers to PCPR generation that stemmed from the non-standardization of pathology reporting or other HIT-related limitations. For example, in all four pilot trials, patients requested the addition of a visual diagram within the PCPRs to help clarify the location of their cancer findings. Functionality to include pictures, though, was not possible in the available EMR technology. The colorectal polyp case study reflected the challenge of combining multiple pathology specimens into a single PCPR. These barriers may continue to lessen as EHR systems advance their capacity for common data elements and other features that increase the consistency of pathology reporting but emphasize the importance of continuous, dedicated HIT engagement with local and national HIT professionals to ensure PCPRs evolve alongside pathology documentation tools.

Maintenance
Our experiences highlighted two areas needed to support the sustainability of PCPRs in practice. First, there is a need for continuous HIT prioritization of PCPR tools. As our colorectal polyp case study demonstrated, lack of HIT prioritization can lead to delays in PCPR development and maintenance over time, especially as HIT environments inevitably change within a given health system. Second, we recognized the need for continuous engagement of multidisciplinary clinical stakeholders (e.g., pathology, oncology, radiology), as guidelines for pathology reporting and cancer staging evolve. Having continuous engagement of clinical stakeholders will work to ensure PCPRs maintain their clinical accuracy and relevance, such as through regular updating of clinical algorithms, which is critical for their sustainment.

Conclusion
There is great potential for PCPRs and other patient-centered tools for cancer care delivery to improve patient satisfaction, knowledge, quality of shared decision-making activities, and ultimately receipt of patient-centered cancer care. Our experiences from these four pilot trials parallel other studies that confronted significant challenges with implementation of patient-centered tools for cancer care, including barriers to clinical complexity, workflow, and use of HIT [10]. These data also demonstrate the need for tools that enhance patient engagement to be tied to opportunities to reduce costs, improve quality, and reduce waste in care delivery systems [19]. To propel these tools forward, coordinated efforts are needed to address multilevel implementation barriers across that limit health system capacity for patient-centered care [11].