Keywords

1 Introduction

Multiple Sclerosis (MS) is an inflammatory demyelinating condition resulting from an autoimmune attack of the myeline sheath of the neurons in the central nervous system [1, 2]. It is a severe neurological disorder with varying presentations and disease progression from person to person [1,2,3]. The global prevalence of MS is increasing, disproportionately affecting productive age groups and females [4]. Highly effective disease modifying treatments are increasingly utilized for this lifelong condition with the ultimate aim of reducing the frequency and severity of relapses, slowing the disease progression, and managing symptoms [5]. Along with non-pharmacological [6] and rehabilitation services [7], the overall goal is to improve quality of life. Time plays a crucial role to achieve the best outcome in MS care [8]. One of the factors delaying the timely intervention is the organization of care processes. There are international clinical guidelines and national efforts to streamline the MS patient pathways [8,9,10], however, these are recommendations prone to limitations in actual implementation. The care processes are described in a text and flowchart formats in many situations. Managing complex conditions such as MS further complicates the already complex organization of healthcare system. When a business processes for non-linear and less predictable processes, such as managing MS, presented in a narrative form, comprehending the complexity becomes difficult.

Visual models are useful in enabling easy understanding of processes, which improves communication and mutual understanding among stakeholders [11]. Several visual modeling languages exist, broadly classified into domain specific (i.e., health or clinical oriented in this case) or general purpose [12]. In this case study, we employed the Customer Journey Modeling Language (CJML), which is designed to visually model both the planned and executed journeys [13]. CJML has previously been used in the healthcare domain and shown to model both the planned journey (or the patient pathway) and the actual journey from the perspective of the patients [14]. With the increasing use of patient centric pathway implementations, CJML is well suited to model patient pathways. In this case study, we explored the potential of CJML in modeling more personalized care pathways from the perspective of service providers.

The healthcare in Norway is organized into municipal and specialized care services [15]. The mandate of establishing diagnoses and initiating treatment for MS resides at the specialist care, usually at the public hospital level, in the department of neurology where MS specialized neurologists are available. Long-term rehabilitation services are mainly provided by a specialized centers for neurological disorders, which is organized independently under the specialist care. Integration of services such as MS along multiple levels of care requires efficient organization of processes. In this study, we emphasize on visualizing the handovers and challenges that follow in the integration of services in the MS patient pathways.

This study is part of larger research project working towards creating a comprehensive toolkit for managing and communicating patient pathways. The aim of this study was to map and visually model the current MS patient pathways within specialized healthcare, including diagnosis, treatment, and follow-up at the hospital level and rehabilitation services at the neurology specialized rehabilitation center from the healthcare workers’ perspectives.

Specific objectives were to:

  • Assess the existing MS care delivery processes at the hospital and the rehabilitation center

  • Identify and map actors and touchpoints in the MS patient pathways

  • Identify practices and procedures related to handovers within and across healthcare levels

  • Assess the feasibility of Customer Journey Modeling Language (CJML) to visually represent the MS patient pathways from healthcare workers’ perspectives

2 Methods

2.1 Study Design, Area and Participants

The study utilized a qualitative study design, using semi-structured interviews to systematically collect insights on the MS patient pathways as seen from the perspectives of healthcare providers’ working within the specialized healthcare (a department of neurology and outpatient clinic and a MS rehabilitation center). This study was conducted in Oslo University Hospital, which serves the largest number of MS patients in the country. The hospital has several specialists and allied professionals dedicated to MS research and clinical care provision. The MS rehabilitation service is situated in one of the neurology related specialized rehabilitation centers close to the city of Oslo. The center provides rehabilitation services to patients with other neurological disorders. The center accepts patients from all over the county for inpatient rehabilitation services. To gain a comprehensive understanding of the MS care delivery process, participants were purposively selected, representing different disciplines, roles, and services within MS care. The final six participants included two neurologists, one health secretary and one nurse working at the hospital MS clinic and a leader and a coordinator (both are health professionals) at the MS rehabilitation center. Half of the participants were females. All participants had several years of experiences in their respective role. All approached healthcare providers agreed to participate in the study. The data collection spanned from June – November 2023.

2.2 Data Collection

The material was collected in two phases; In the first phase, a semi-structed interview guide was used to gather detailed information about the touchpoints and interactions the participant (e.g., the nurse) was involved in during the MS patient pathway. The interview guide aimed at exploring step-by-step organizational and clinical processes undertaken by the participants and their colleagues in their respective roles. These spanned from the very first contact with the patient (e.g., receiving a referral from the general practitioner) throughout the entire treatment/care provision up to patients’ exit. The interview guide was semi-structured with open-ended questions to facilitate the free narration of the care processes. It was developed by the researchers and iteratively improved throughout the two phases of the interviews.

Based on the insights from the first interview, the first author (BB) visualized the MS patient pathways using Customer Journey Modeling Language (CJML). The visualizations were then used as discussion points in the following round of interviews (second phase). The aims of these post-modeling interviews were to present the visualizations and get feedback on 1) potential gaps from the initial round of interviews, 2) errors and needs for adjustment, 3) understandability of the visualizations and 4) the usefulness of these kinds of visualizations. The final visual models were based on iterative improvements in addition to the feedbacks from the participants.

Except one virtual meeting, the data collections were conducted physically, at the premises of the institutions. The interviews were conducted by two of the authors (BB and IKLS) who are not associated to the MS patient care, and while one lead the interview the other took notes. Each interview lasted for around 1,5–2 h in each phase. Since this is for process map** and modeling, we did not aim for information saturation to determine the adequacy of number of participants, rather diversity of participants. Most of the interviews were conducted in English. However, two of the interviews were conducted in local language with notes later being translated into English. All interviews were recorded and later transcribed.

The study was approved by regional and institutional ethics committee and collected in accordance with the Helsinki Declaration [16]. All participants signed informed consent before participation. Data was stored in a secure directory and results presented anonymously, by observing the personal data protection rules [17].

2.3 Data Analysis

In the first step of analysis, two of the authors (BB and IKLS; both external researchers) listened to the recordings and took notes, supplementing the original notes taken at the time of the interviews. Then the entire MS patient pathways in both institutions were divided into sub-processes for the sake of simplicity during modeling procedure and presentation of the CJML models. The notes were then sorted based on these sub-processes for further analysis of the interview materials. The next step in the analysis process was that the first author (BB) identified and mapped actors (patients, individual healthcare professionals, or a unit in healthcare system) and touchpoints (i.e., actions, medium, and instances of communications between the patient and the healthcare institution actors), and further visualized these using CJML. The visualizations were then used in the follow-up interview with each participant to get feedback, ensuring that the insights and details were understood correctly, while also providing an opportunity for additional data and details left out from first interview. The interview material from second phase interviews also were sorted into sub-processes in the same manner as the first phase interviews.

The data material was revisited to identify common themes discussed during the two phased interviews. Using the sorted data under each sub-process, the first author grouped and labelled the major themes discussed by the participants. Main themes are presented together with illustrating quotes, ensuring transparency in the interpretation of results. Thus, several rounds of revisiting the material were required to iteratively improve the final version of the model. The output of the process, visual models of selected sub-processes using CJML is presented to illustrate parts of the current MS patient pathways in these institutions.

3 Results

The result section is structed into sub-processes in the MS patient pathway and are hence divided into the following sections: 1) referrals, 2) internal handovers, 3) admission processing, 4) diagnosis, 4) treatment and symptom management, and 5) follow-up. As most sub-processes apply to both institutions (i.e., the hospital and the rehabilitation center), findings are presented together to allow for comparisons.

3.1 Referral Processing

There are two main routes of referrals to the hospital for a MS specialist care: 1) directly from primary care, mainly from the general practitioner’s office presenting with mild or moderate symptoms, or 2) referred from the same hospital or other healthcare outlets after receiving emergency management. Figure 1 shows the CJML model of the referral processing sub-process using a swimlane diagram. Similarly, the rehabilitation center also receives referrals from a MS specialist at the hospital or general practitioner through regional approving authority (Fig. 2). For both facilities, the models in Fig. 1 and 2 show the first referral routes to illustrate the involved actors and touchpoints. The other route is not shown due to space restrictions.

Assumptions:

Referral is from GP for a mild or moderate symptoms.

All the requests from the communicator are accepted; or not declined by the receiver of the request or information.

Fig. 1.
figure 1

CJML model of MS patient pathway sub-process at hospital specialist care

Full size image
Fig. 2.
figure 2

CJML model of MS patient pathway sub-process at the specialist rehabilitation center

Full size image

Major themes discussed by the participants regarding this sub-process are as follows.

Communication Between Institutions

The participants discussed the ways in which the referring professionals communicate to their institutions. Some made a call before processing the referral requests, while others only referred the patients through the digital communication channels. The content of the referral letter also varied. The participant from rehabilitation center stated that “some time the referral letter says, ‘this patient needs rehabilitation’. Others (mentioned a professional by name) detail what specifically the patient needs the rehabilitation service.”

The participants mentioned that the speed at which the referral is handled by the receiving institution is commonly within 10 days. The waiting time after acceptance of the referrals varied because of the patient load and occupancy. This is only for the referrals between institutions; the internal handovers are processed differently. This shows variabilities in the quality of referral communication.

Information Communication Tools Usage

Each institutions used different vendors for the electronic patient record system, which lead to additional communication touchpoint to obtain the needed patient information. The use of Silo electronic recording systems discussed as one of the information flow hinderers between facilities. The effect of using similar, or at least interoperable systems, would mean real time access of patient information by all actors involved. Participants also mentioned the ongoing efforts to improve the interoperability with limited implementation and it remains a challenge.

3.2 Internal Handovers

In addition to the referrals between institutions tasked with different care provisions, there is internal handovers for semi-independent actors in the MS pathways at the hospital specialist care. The neurologist who diagnosed the patient hands over to the MS nurse at the out-patient clinic to administer the treatment and manage the patient (Fig. 5). There are differences following the type of medication to be administered, however, the majority who take intravenous infusions are managed, both clinically and organizationally, by the MS nurses at the out-patient clinic with limited or no involvement from the diagnosing MS specialist neurologist.

Informal Communication Works Best Internally

The actor who is handing over the care internally communicates to the other party, often in person despite the presence of internal electronic communications channel. The digital communication is also used but informal communications believed to be feasible in small institutions. Any communication that transcends a department, for example, to a laboratory department, however, follows standard operating procedure of the hospital.

Information Communication Tools Usage

Internally, the same electronic tool is being used. This allowed a real time information exchange (including yellow notes for urgent attention) between the neurologists and MS nurses. The electronic health record under use required some expertise to locate the information a given professional may need, which might have hindered the effective use of the digital tool for internal communication. The administrative information is easier for the health secretary to obtain and process but maybe challenging for the clinicians. Health secretary said that “you have to click several places to retrieve information, which is easier for us since we use the system a lot.”

3.3 Admission

This sub-process applies to the rehabilitation services. However, the information needs, and retrieval applies to the hospital services to some extent, except that majority of the patient data about the MS is generated at the hospital level. Coupled with the less detailed description in the referral letter, the retrieval of information during and after admission became very laborious, which can also lead to inaccuracies. The participants stated that the rehabilitation center conducts a lot of communications via different channels to obtain patient specific information needed to commence the service. One of the participants said, “we do a lot of calls…does not matter who makes the calls as long as the information is obtained”. Figure 3 shows the CJML model of the communication needs that arise although it is not possible to list all the actors as it depends on the individual cases.

Flexible to Accommodate More Patients and Operate More Quickly

Both the health secretaries who manage the appointment scheduling and the physicians cooperate to help the patients get appointments faster. The patients are also given options to get the diagnosis at private specialist practitioners if there is no possibility of finding appointments shortly. The health secretary said that “We also collaborate with the private sector to give the patients options to get timely diagnosis if we could not find free spots with our specialists”. Clinicians also use their spare time and adjust their schedules to accommodate patients. One of the physicians said that “We try to exploit the system or adjust somethings…and I think we are actually quite good at adapting.”

Fig. 3.
figure 3

CJML model of patient and stakeholder communication to retrieve patient information at the rehabilitation center.

Full size image

3.4 Establishing Diagnosis

The diagnosis for MS is established by the neurologist at the hospital specialist care. The process involves imaging and diagnostic lumbar puncture, which involves other actors in the pathway. All the internal communications between actors were made via the same EHR system, in addition, in person consultation of senior physicians. Figure 4 demonstrates the actors involved with the communication touchpoints.

Fig. 4.
figure 4

CJML model of MS patient pathway diagnosis sub-process at the hospital

Full size image

Variable Approaches Among Clinicians

Some clinicians take the initiative to contact their patients to notify the results. One participant stated “I think there might be a difference there. I usually call the patients after the MRI because I know that they are very anxiously waiting”.

Uncertain Waiting Time for Diagnostic Results

Due to waiting time for the laboratory and imaging services, the diagnostic results might not come in the duration the neurologist hopes to receive. This affects the next scheduling and patients are anxiously waiting for the result. For emergency cases, however, the neurologist can put a flag in the electronic order so that the result turnover is quicker.

3.5 Treatment and Symptoms Monitoring

This sub-process at the hospital and at the rehabilitation center aimed at different focus, but the actors’ involvement at the team level makes the organization of care somehow similar. First, the treatment is administered by the out-patient clinic (polyclinic) which operate semi-independently. The patient communication, internal communication between departments and the diagnosing neurologist, and the monitoring of treatment progress is planned and managed after the handover is completed.

Fig. 5.
figure 5

CJML model for infusion treatment at the neurology out-patient clinic

Full size image

3.6 Follow-Ups

Since the MS management is a lifelong process, the follow up at the different service outlets is warranted. Specially, patients are scheduled for a follow up with preferably the diagnosing neurologist at least after 6 months of diagnosis, unless there was no other indication for more frequent visits. The neurologists promise their patient an estimated time for the follow up visit but are not in control because of the long waiting list. The scheduling is also monitored by the health secretary who has to follow the queue in the waiting list. Sometimes, the patient has to be scheduled with another neurologist than the one who diagnosed them first, which is not welcomed by the neurologists.

4 Discussion

Employing an iterative approach, we visually modeled the MS patient pathways at the hospital specialist care and a rehabilitation center using CJML. The three main contributions of this study were identifying and dealing with the challenges and opportunities of visually illustrating care processes with highly personalized care, as in the case of MS patient pathways; applying CJML to model from the perspective of the healthcare service providers; and the complexities of modeling handovers by including more than one care level. In the following paragraphs, we discuss these points selectively to share our approaches. The process of visually modeling a care process presents opportunities to bottleneck analysis for the future improvements of services. This study contributed to the first stages of bottleneck analysis and process optimization [18]. Assessing and modeling an existing care process, the ‘as-is’, is a step in develo** a ‘to-be’, or future ideal organization of care by addressing the bottlenecks or even optimizing a functioning pathways [19, 20]. The studied MS patient pathways is based on international guidelines and national care process organization recommendations [8, 9]. Assessments like this, along with insights from studies such as [21] on care variations among patients, offer valuable input to develop integrated MS patient pathways, centered around a life-course approach, based on a proactive, multidisciplinary, and patient pathways concept.

Visually modeling every scenario for health condition with a variable presentation and disease course is a difficult task [11]. As to the first contribution, we understood the necessary variations arising from individualization or other clinical factors that can be modeled separately. However, we only modeled the commonest pathways while recognizing individual variations within. For example, there are several treatment options for MS tailored to individual factors [22]. In the study area, however, majority of the patients are using highly effective disease modifying therapy, which is modeled as a common treatment pathway on a higher level of abstraction. Our modeling does not go into the clinical decisions where further individualization of the therapy is made [22]. Therefore, the abstraction of a visual model is at a higher or group level. This may have reduced the usual trap of a complex visual models of a process that is hard to understand and often not used [11]. Modeling should always allow individual variations because of unpredictable factors in the care delivery process [23]. Such complexity calls for a more adaptive modeling that has a flexibility to present the individualization process in a planned pathway.

Since patient care for conditions such as MS spans over lifetime and involves several stakeholders, considering the integration of care throughout multiple levels of healthcare is vital [24, 25]. Where there is inter-institutional communication, the flow of patient information is affected by several factors. We identified challenges encountered during referral processing. One of the challenges emanates from the usage of a silo electronic health record (EHR) systems, placing additional workflow issues, where providers spend time in collecting information about the patient, otherwise would have been automatically retrieved [26]. In addition to the logistics it requires, the quality of patient information is compromised when other ways of communications were employed. During modeling, we also run into an overwhelming number of actors and touchpoints, which was also different for individual cases. Pertaining to modeling the handovers contribution of this study, the work around in the modeling was to provide assumptions (Fig. 3). The within institution handovers where the same EMR is used have not encountered as much difficulty in finding patient information as in the case of referrals between institutions (Fig. 5). One important aspect discussed in the process of handovers and management of follow-ups was expressed by the physicians, concerning the long waiting time, which is against the very idea of ‘time matters in MS’[8].

Visual modeling of a process works best for predictable and linear processes [27, 28]. Most business process modeling languages work well when a process meets those criteria. The healthcare domain, due to its complex nature (e.g., dealing with human health, not a production line), has been a challenging area to easily adapt process modeling languages and apply [12]. Therefore, several studies demonstrated that the domain requirements are being met by different approaches, including adapting a modeling language to the domain needs [29]. Push factors such as using a patient-centered approach in a care delivery model, with the growing literature on patient journey studies pave the way for opting for a customer-oriented modeling languages over process modeling languages with a focus on beyond line of visibility for customers (service blueprints) [14]. One such example is CJML [13]. In this study, we demonstrated how information gathered from healthcare works with multidisciplinary background and roles can also be used to model. The CJML provides two diagram types, where swimlane is one of them. We used swimlane diagram type where all the actors, including the patient can be depicted (as presented in the models in Fig. 1, 2, 3, 4 and 5) in a model. Processes happening beyond the line of visibility for the patient can be modeled and used as a service blueprint best using other modeling languages. Approaching from the client perspective provides opportunity to model the individual pathways whereas modeling from the perspective of the service providers presents a holistic process of the organization of a given care. The advantage of approaching from both providers’ and patients’ perspectives for a more standardized patient pathways is underway within the umbrella project this study is a part of.

This study has strength and limitations as any research projects. We worked with the same participant twice –before and after the modeling exercise. This approach, we believe, allowed us to refine the modeling process. There were several questions that we identified in the process of modeling the first iteration based on the first interview, which we used to get clarification in the post-modeling interview. Without such a process, there would have been incomplete or misleading visualization of the process. The inclusion of professionals with different roles and responsibilities in the patient pathway also provided a comprehensive map** of the processes.

Using the patient-centered modeling language helps to put the patient at the center of the process modeling even though the data is from the providers’ perspectives. Commonly used visual models present the service blueprints, dealing with the hidden part of the process from the users, missing the crucial touchpoints with the patients along with the end-users’ feelings that affects the overall effect of care delivery process [13].

Since creating personas or grou** cases is a challenging task as the care provision is highly personalized and the touchpoints are dependent on the care providers’ innovative approach and dedication, we presented the models with assumptions by only modeling the commonest, yet without going into low level of abstractions. Such categorization cannot be exhaustive enough to include all the different variations. During the interview, we explored ways of categorizing and creating personas with the participants. The lesson learned from the process was that it is challenging and probably needs a refined methodology for future studies of similar kinds.

Future similar studies may consider a hands-on modeling of the process with the participants. We only involved them in the refinement process to some extent but co-creating the models could give them more agency in presenting the process even more accurately. It has been demonstrated earlier that non-modelers can easily adopt CJML and make precise models of a process [30]. Especially when there is an attempt to create a ‘to-be’ type of patient pathways, hands-on approach might be even more relevant. We also did not carry out a thorough feasibility study by involving more stakeholders than the participants in the interview. This can be improved by including other healthcare workers to thoroughly assess the understandability, usability, and feasibility of CJML models.

5 Conclusions

The process of modeling a care system allows for reflection on care delivery organization. Map** the actors and touchpoints in MS patient pathways at a hospital and specialized rehabilitation center, involving healthcare workers in various roles, helps comprehend the current organization of care processes. Visualizing highly individualized patient pathways for chronic conditions managed across multiple care levels with numerous stakeholders is inherently complex. While modeling languages like CJML can aid in visually representing patient pathways, further studies are needed to enhance methods and address all domain needs comprehensively.