We used qualitative methods to develop the CDK4/6 inhibitor real-world patient information tool. The development process consisted of five steps: (1) establishment of a multidisciplinary steering group, (2) mapping of the patient journey and patients’ needs through focus groups and semistructured interviews, (3) extraction of real-world outcome data from electronic health records systems, (4) prototyping of the tool, and (5) pilot evaluation with the targeted patient population using semistructured interviews (Figure 1).

The study protocol was approved by the Medical Research Ethics Committees United (Nieuwegein, the Netherlands; reference SDB2021-010). Ethics and research governance approval were obtained from the local medical ethics committees of the participating hospitals. Participating patients provided written informed consent.

The development of the information tool was initiated by Santeon, a Dutch hospital group in which 7 top clinical hospitals openly collaborate to improve medical care [21]. The project initiators (postdoctoral researcher MQNH, program manager CFvUK, and hospital pharmacist EMWvdG) established a multidisciplinary steering group for the development of the tool.

The patient journey mapping (ie, inventory of patients’ experiences from diagnosis until disease progression) served as the starting point in the development of the information tool. The goal was to obtain insight into patients’ experiences from diverse patients to help identify information and support needs that needed to be addressed in the patient information tool. We collaborated with the Patient Journey Lab, a company specialized in patient journey mapping, to develop the patient journey. First, to roughly outline the patient journey, 2 focus groups were conducted by MQNH and an independent research consultant (experienced in conducting focus groups) with 5 and 4 patients, respectively, that lasted about an hour. They focused on the following topics: (1) information provision, (2) decision-making to start the CDK4/6 inhibitor, (3) unmet information needs, and (4) perspectives on the use of real-world outcome data. Second, to verify and fill out the patient journey in greater detail, semistructured interviews, with an expected duration of 30 minutes, were conducted by MQNH with 3 patients and the 2 oncologists and 3 nurse specialists who participated in the steering group. All recordings were transcribed and summarized by MQNH. Manual summative content analysis [23] was used to infer meaning from recurring keywords and messages. Based on the topic list, keywords were selected to highlight in the text to count their frequencies before defining and contextualizing them in our final summary.

To inform new patients about the treatment, the information tool includes outcome data of 229 patients treated between 2016 and 2020 in 6 Santeon hospitals (St Antonius Hospital Utrecht/Nieuwegein, Martini Hospital Groningen, Canisius-Wilhelmina Hospital Nijmegen, Catharina Hospital Eindhoven, OLVG Amsterdam, and Maasstad Hospital Rotterdam). More detailed information on the real-world cohort and the data collection has been reported previously and is available elsewhere [24,25]. Briefly, the data collected included PFS, the reason for ending treatment, the severity, frequency, and timing of all side effects and toxicities (with a special focus on neutropenia), and the timing and frequency of all treatment modification strategies. They were derived from routinely collected electronic health records, with 1 hospital excluded due to incomplete documentation of palbociclib use. To ensure the robustness of the real-world data to be included in the tool, we compared in this dataset outcomes in a cohort of 229 patients with HR_pos, HER2_neg metastatic BC treated with palbociclib plus fulvestrant in second- or later-line therapy to those reported in the pivotal PALOMA-3 trial that formed the basis for regulatory approval of palbociclib in this setting. On the basis of the comparison, the expert work group concluded that given the greater clinical heterogeneity and the mismatch in toxicity profiles compared with trial participants, real-world outcome estimates were more representative for informing patients and were therefore used in the information tool. [24,25]

In consultation with the steering group and based on the patient journey mapping, it was decided to develop a digital information tool to allow future efforts to link the tool to patients’ electronic health records. Such a linkage provides two main benefits: (1) it gives patients access to all medical information in one place, and (2) it allows the tool to display up-to-date real-world outcome data based on automatic updates retrieved from personal data stored in electronic health record systems. To examine how the tool could best be integrated in the digital landscape of the Santeon hospitals, the Patient Journey Lab inventoried the software potential per hospital, including available health applications. As CAREanimations offers accessible information suitable for most people, including those with low (health) literacy, tailored to each step in the patient journey, it was selected as the development and implementation partner [26]. In collaboration with CAREanimations, the initiators developed several low-fidelity prototypes, which were reviewed by the steering group to arrive at a high-fidelity prototype for pilot-testing.

We carried out a pilot evaluation in 2 steps. First, before implementing the information tool, we assessed the status quo, that is, patients’ experiences with treatment and their met and unmet information and support needs. We did this to validate the findings of our previous unmet need assessment. Second, we assessed patients’ views on the usefulness and acceptability of the high-fidelity prototype of the information tool and inventoried suggestions for improvement.

In the first step, we conducted in-depth semistructured interviews with women with HR_posHER2_neg metastatic BC who (1) were using or had stopped using the CDK4/6 inhibitor Palbociclib within the last 3 months, (2) had sufficient command of the Dutch language, (3) were 18 years or older, (4) did not have (severe) cognitive impairment that would preclude them from providing informed consent, and (5) had not been involved in the development process. We obtained insights into patients’ (1) experiences with information provision about their diagnosis, prognosis, and treatment with CDK4/6 inhibitor; (2) experiences with treatment (eg, side effects and impact on daily functioning); (3) experiences with treatment decision-making; and (4) preferences regarding receipt of outcome data. In addition, we obtained information about sociodemographic characteristics (age and educational level) and medical history (date of primary BC diagnosis, treatment history; refer to Multimedia Appendix 1 for the full interview guide). We recruited participants from 5 Santeon hospitals (ie, St Antonius Hospital [Utrecht/Nieuwegein], Martini Hospital [Groningen], Canisius-Wilhelmina Hospital [Nijmegen], Catharina Hospital [Eindhoven], and Maasstad Hospital [Rotterdam]). Patients were invited to participate by their oncologist or nurse specialist. The interviews were conducted by EGE (researcher with interviewing experience).

In the second step, we evaluated the prototype of the information tool through individual semistructured interviews with a new cohort of women not previously involved in the development or pilot-testing, meeting the same inclusion criteria as described in step 1. Participants were recruited at the same 5 Santeon hospitals and via the Dutch Breast Cancer Society. Recruitment through the Dutch Breast Cancer Society broadened the target population, as participants were also treated at other hospitals, which accelerated the recruitment process. Two weeks before the interview, participants were provided with access to all the animation videos contained in the information tool and the graphical information intended for use during the consultation with their oncologist. They were asked to view all the materials and take notes of any suggestions for possible improvements. After they reviewed the information, a telephone interview was conducted by MQNH or EGE. The interview guide to assess patients’ experiences with treatment and their met and unmet information and support needs was used and extended to also obtain insight into patients’ views on the completeness and clarity of the CDK4/6 inhibitor information tool and to collect suggestions for improvement (refer to Multimedia Appendix 1 for the full interview guide).

All interviews (in the first and second steps) were audio-recorded with the participant’s permission and transcribed. The transcripts were analyzed using a combined technique of deductive and inductive thematic analysis [27,28]. For the deductive approach, we constructed an analytic framework based on predefined themes in the interview guide. In addition, we conducted open coding (inductive) on the transcripts to allow the inclusion of other relevant themes. We generated categories and abstractions through detailed review, quantifying keywords and phrases, and comparing findings. As educational level was considered to be an important predictor of patients’ comprehension of and satisfaction with the information tool, we explored differences in participants’ responses based on their educational background. All transcripts were coded by 2 researchers (MQNH and EGE) independently. Discrepancies were resolved through consensus. Qualitative data analyses were performed using Atlas.ti (version 9; Lumivero) software. Quotes are provided to illustrate our findings.

The steering group consisted of patient representatives or advocates from the Dutch Breast Cancer Society (n=3), oncologists (n=2), specialist nurses (n=3), an information architect (n=1), and a health information consultant (n=1; for more details, refer to Multimedia Appendix 2). The steering group convened for 4 cocreation sessions. Figure 2 provides an overview of the focus of each session. The steering group was involved throughout the development process.

The findings from interviews and focus groups with patients (n=12) and clinicians (n=5) were used to map the patient journey of women undergoing treatment with a CDK4/6 inhibitor, from the start of treatment to its discontinuation (Table 1). A detailed overview of the patient journey is provided in Figure 3. Findings from the interviews with clinicians about the requirements for the information tool were in line with the findings from the patient interviews. All participants valued real-world outcome data in addition to clinical trial results, preferably displayed in a leaflet or app (connected to the electronic medical record). Information needed to be tailored to each step of the patient journey. Repetition and gradually offering more information were believed to be key. Even though dealing with a wide range and likely negative outcomes may be complex, participants wished insight into the timing, frequency, and severity of side effects, as well as the timing and reasons for ending treatment, so they could compare their outcomes to those of peers. Most patients wished to be better informed about possible medication alterations. Information on what type of treatment CDK4/6 inhibitor is, as well as what to do when experiencing side effects, was also valued. A summary of the key findings from the focus groups and interviews conducted to map the patient journey can be found in Table 2.

Real-world outcome data were extracted from a cohort of patients treated with palbociclib in combination with fulvestrant in routine clinical practice. In this cohort, the median PFS was 11.6 (95% CI 10.2-13.9) months, which was longer than the 9.5 (95% CI 9.2-11.0) months reported for the PALOMA-3 trial population (more detailed information is available in the study by Hackert et al [24]). This difference may partly reflect differences in outcome assessment between clinical trials and routine clinical practice. When the real-world cohort was stratified according to PALOMA-3 eligibility criteria, differences in outcomes were observed. Patients who would have met the PALOMA-3 eligibility criteria had a median PFS of 14.1 (95% CI 11.5-17.0) months, whereas patients who would not have met these criteria had a median PFS of 10.2 (95% CI 8.8-12.3) months.

The occurrence of hematological toxicities and nonhematological side effects associated with palbociclib was generally similar in real-world settings compared with the clinical trial data. However, patients in routine clinical practice reported certain side effects more frequently, particularly fatigue and decreased appetite. In clinical practice, 54% (124/229) of patients experienced at least 1 cycle delay, 39% (89/229) had at least 1 dose reduction, and 26% (59/229) had at least 1 dose interruption, compared with 36% (124/345), 34% (117/345), and 54% (186/345), respectively, in the PALOMA-3 trial. Detailed findings on patient-reported experiences are described in the study by Vaisson et al [22].

To support the development of the information tool, the extracted real-world outcomes were summarized using lay terminology and reviewed together with the steering group. Based on the extracted outcome data, decisions were made in collaboration with the steering group regarding which outcomes to include in the tool and how these should be presented to patients. Particular attention was given to ensuring that the translation of clinical outcomes into patient-friendly information accurately reflected the limitations of real-world data, such as potential selection bias, missing or incomplete data, and variability in outcome measurement and reporting. For example, rather than presenting outcome distributions based solely on real-world data, outcomes from clinical trial data and real-world data were compared, and presentation choices were made based on this comparison. This approach aimed to provide balanced and transparent information for patients using the information tool.

Based on the requirements set by the steering group (using the results from the patient journey mapping and the patient and clinician needs assessment), prototypes for the information tool were developed. Figures 4 and 5 provide a summary of the patient journey, including components of the information tool, to support information provision and treatment decision-making at each phase of the patient journey. The information tool consists of three main components: (1) KIJKgesprek (Stichting Kijksluiter), (2) KIJKsluiter (Stichting Kijksluiter), and (3) KIJKbericht. The 3 components are described in Figure 6. The steering group reviewed the prototypes, and in an iterative process, the prototypes were refined to a version ready for pilot-testing—the high-fidelity prototype.

Overall, step 1 showed substantial gaps in patients’ understanding of CDK4/6 inhibitor therapy, wide variation in information needs and preferences, and challenges in coping with treatment burden and decision-making. These findings underscore the need for structured, accessible, and patient-tailored support materials that incorporate realistic expectations of treatment and side effects. Step 2 demonstrated that the prototype information tool addressed many of these needs. Even though preferences for receiving outcome data varied between patients, many participants did find this information useful. Together, these insights support the incorporation of real-world outcome estimates into the tool.