Dual-Perspective Modeling of Patient Pathways: A Case Study on Kidney Cancer

Patient pathway has become a key concept in the organization of healthcare, intended to address issues pertaining to coherency, seamlessness, and accountability in a system with increasing complexity [1, 2]. Understanding and representing patient pathways is essential in healthcare, as it helps identify potential bottlenecks and informs targeted interventions to improve patient care [3]. Moreover, while patient pathways are mainly used within hospitals, their use can also strengthen coordination between healthcare actors. A related concept is the patient journey, which takes the perspective of the patient throughout the pathway. Patient journey mapping is increasingly being adopted in healthcare settings to provide insights into the patient experience and to support communication with patients [4].

For simplicity, in the following, we use the term ‘patient pathway’ to describe pathways as designed by the healthcare system and the term ‘patient journey’ to describe actual, individual patient encounters throughout an illness period. We, however, maintain a patient-centered perspective for both. While many definitions of patient pathways suggest some level of patient-centered perspective [5], pathways tend to focus on clinical guidelines and other work processes behind the line of visibility of the patient [6]. As patient pathways involve numerous stakeholders and intricate decision-making processes, selecting a modeling approach that effectively captures both patient-provider interactions and the coordination of healthcare services is crucial to ensure the delivery of patient-centered care.

The complexity of healthcare processes, along with the need for effective communication and production planning, has led to a growing interest in using modeling languages to represent patient pathways. Several modeling languages have been used, including Unified Modeling Language [7], Business Process Modeling Language [8], extensions of such languages [9], Customer Journey Modeling Language [10], and a multitude of less formal patient journey maps [4], each with their strengths and weaknesses. The use of modeling languages for documenting patient pathways can reduce variability, help facilitate interdisciplinary collaboration, streamline decision-making processes, and ultimately improve patient outcomes [2, 5].

In this study we focus on the feasibility of the Customer Journey Modeling Language (CJML) for modeling of patient pathways, as planned by the healthcare system, and patient journeys, as experienced by individual patients. Offering a vocabulary, a metamodel, and purpose-specific diagrams, CJML is equipped with journey-specific constructs including touchpoints, actors, channels, phases, and user experiences [11]. CJML’s patient-centric approach makes it particularly relevant for modeling patient pathways in healthcare. However, CJML may need adaptation to address the complexity and specific nuances of healthcare processes, as well as the interactions between healthcare institutions and personnel. Hence, there is a need for more in-depth research on CJML in healthcare contexts, including how it may be adapted and improved to suit the specificities of healthcare contexts.

Using kidney cancer as a case study, we explore how patient pathways and journeys can be identified from a dual perspective (both healthcare personnel and patients). We investigate how CJML can be utilized to visualize and compare these pathways and journeys. Additionally, we examine its application in healthcare settings, providing in-depth insights into the pathways and journeys from both perspectives, thereby enabling comparison. In specific, we address the following research questions:

To address these research questions, we present a case study of kidney cancer care conducted by health service researchers and medical doctors at a major university hospital in Norway. Kidney cancer typically develops slowly and is often detected incidentally during investigations for other conditions [12]. The involved hospital treats approximately 100 kidney cancer patients annually. The vast majority of these patients undergo either partial or radical nephrectomy. This study aims to contribute to the ongoing discourse on the application of modeling languages in healthcare and emphasize the importance of patient-centered approaches in representing and understanding patient pathways.

The study adopted a multi-method approach dominated by qualitative research methodology. The first phase of the work was conducted in two parallel streams: 1. Detailed insights into patient pathways (healthcare’s perspective) and 2. A longitudinal study of patient journeys (patient’s perspective), see Fig. 1. Based on this, we modelled the kidney cancer pathway and patient journey using CJML. In the second phase the feasibility of CJML in healthcare settings was evaluated.

In this section, the results of our study are presented. First, the kidney cancer patient pathway, as seen from the hospital perspective, is described and visualized. Then, an actual patient journey, as seen from the patient perspective, is presented, before the two perspectives are compared. Finally, findings from the feasibility interviews are presented, exploring the use of CJML-based visualizations in healthcare settings.

In this study, we utilized CJML to model a kidney cancer patient pathway from observations, workshops, and interviews with healthcare personnel, as well as a detailed patient journey of one patient from interviews and a longitudinal diary study. Moreover, we collected feedback from urologists and coordinators on the feasibility of using CJML modeled pathways in practical healthcare settings. Below, we discuss the key findings related to our three research questions. First, we discuss the effectiveness of our methods for precisely identifying and modeling patient pathways and journeys. Second, we discuss the insights provided by applying a dual perspective. Third, we discuss the feasibility of CJML models in healthcare settings.

Mapping the Patient Perspective. The longitudinal study offered a detailed view of the patient journey, capturing many of the touchpoints that patients experience. We find that the diary method, where patients log each new touchpoint, is well-suited as a basis for modeling patient journeys. However, the diaries alone are not sufficient to provide a full picture of the patient journey, as patients tend to forget to log some touchpoints, such as appointment notifications or phone calls with healthcare personnel. Therefore, interviews are essential to supplement and clarify diary entries, gather patients’ assessment of touchpoints, uncover missed entries, and hence, to provide a fuller picture of the patient journey. For the patient journey detailed in this paper, about eighty percent of the touchpoints were logged in the diary, with the remaining twenty percent identified during interviews. This is consistent with previous findings, which suggest that informants typically report between 50–70% of touchpoints [16]. The diary method enriches the toolkit for collecting patient journey data, complementing focus groups (i.e., [17, 18]), surveys [19, 20], document analysis [21], and interviews (i.e., [22, 23]), by facilitating more detailed patient journey maps. Of the data collection methods available, the diary method ranks among the most detailed. However, the challenge with the method is that it is resource intensive, and there is a risk that participants drop out during data collection. Despite these challenges, our study reaffirms the method’s suitability for accurately modeling patient journeys with CJML, thus providing healthcare personnel with an overview of what a patient journey may look like from the patient perspective, which they today largely seem to lack.

Relatedly, the method provides insights into patients’ experiences with each touchpoint, which can be instrumental in identifying bottlenecks and improving healthcare services overall. An example of the metrics and presentation of patient experience can be found in [13]. By gaining knowledge on patient experiences throughout the patient journey and in relation to specific touchpoints, we obtain a richer, more detailed, and more dynamic understanding of patient experience as it evolves throughout a patient journey, compared to the snapshot provided by measuring patient experience at a single point during or after an illness period. Therefore, our method may be a valuable supplement to patient-reported experience measures (PREMs) [24]. The specificity of the measures entails that they capture particular challenges that patients face, making them well-suited as basis for improving healthcare services. Moreover, by offering insights into experiences over time, researchers can more accurately capture how the overall patient experience is shaped, and which factors that are most important to patients.

Mapping the Patient Pathway. Our study highlights the necessity of a iterative process to accurately map the patient pathway. This includes recurring workshops with healthcare personnel for data gathering and drafting initial pathway sketches, followed by meetings to review initial visualizations, where healthcare personnel provide feedback on the pathway’s accuracy and completeness. Patient pathways are complex, and carrying out data collection in steps ensures that the data and subsequent visualizations are as comprehensive and accurate as possible. This approach aligns with current literature on the development of pathways [20].

Feasibility. Our findings underline how the usefulness of CJML visualizations depends on the context. The results from interviews suggest that while healthcare personnel may not find such diagrams particularly useful in their everyday work, such models may be useful as organizational, instructional, managerial, and communicative tools, such as to map practices across roles, arrive at a common understanding, identify efficiency potentials, and communicate these to stakeholders, as well as in training and educational settings. Relatedly, the findings illustrate that the appropriate level of detail in visualizing patient pathways depends on the purpose and the intended recipient. For example, the appropriate level of detail is likely higher when representations are meant to provide an overview and identify bottlenecks within or across departments than when visualizations, for example, are meant to communicate such bottlenecks and improvement potentials to hospital managers or health authorities. Moreover, if used in training or for educational purposes, very detailed visualizations may help healthcare personnel or students understand the complexity of patient pathways. Less detailed visualizations may be better suited if the goal is to provide an overview of the key steps in a planned patient pathway at a specific department or hospital. Hence, representations of pathways and journeys should have a clearly stated aim, with visualizations being adapted to the aim and the recipient. However, to get an overview of and be able to make relevant representations in different contexts, a detailed mapping of patient pathways would nevertheless be a prerequisite. In other words, all touchpoints may be important, but their importance in visual representations will depend on the context and purpose. This indicate a need to further formalize the modeling language so that, for example, touchpoints representing reminders may easily be omitted from visualizations.

Theoretical and Practical Contributions. Our results contribute to theory by highlighting the need for duality in healthcare systems modeling. Modern healthcare production systems are often designed largely from the production perspective [25], and although the importance of the patient perspective is often underlined, its role in the design of the service production system is often still small. Our results highlight the critical role of the patient in the design of healthcare systems, especially in long-term care processes. Considering a person with kidney cancer lives with the disease 24/7, how can we design a functional care system purely from the perspective of the actor who meets the person for a few hours every year? Practically, we contribute a concrete tool and example of how such a dual pathway model can be built and used to inform decision-making in patient processes. CJML is an openly available modeling tool with support resources available for healthcare professionals to make use of in their process improvement efforts.

In this paper we have examined the identification and modeling of patient pathways and journeys, the insights gained from applying a dual perspective, and the feasibility of CJML-based models in healthcare. The methodological approach, and the combined exploration of the healthcare- and patient perspective highlight the complexity of patient pathways and patient journeys, illustrating that visualizations can provide a valuable overview of this complexity. Although healthcare professionals may not find these visualizations crucial for daily tasks, our findings suggest that the methods and subsequent visualizations may be useful as organizational, instructional, managerial, and communicative tools.

The focus on one illness, a few patients, and a single hospital department allowed for an in-depth analysis of patient pathways and journeys. This approach has allowed us to explore how pathways and journeys can be precisely captured and modelled, how they compare, and the feasibility of CJML models in healthcare settings. However, the focus on a single illness from the perspective of a relatively limited number of healthcare professionals and patients is also a limitation. The insights provided in this paper should be explored further and tested in other contexts, with other illnesses, and on larger samples. We maintain that our findings related to identifying and modeling patient pathways/journeys, and the feasibility of using these models in healthcare settings have value beyond kidney cancer care. Nevertheless, kidney cancer patient pathways may be relatively short and standardized compared to other illnesses, such as chronic diseases or other types of cancer, indicating a need for further studies focusing on other illnesses. In terms of feasibility, additional studies are needed to better understand the specific contexts in which visualizations may be useful and the appropriate level of detail for presenting patient pathways/journeys to different actors, such as patients, hospital administration, nurses, doctors, or even students. Relatedly, since the findings indicate a need for filtering touchpoints and providing visualizations at different levels of abstraction, future studies should further explore how the formalism of CJML can be extended to more effectively support the healthcare context.