The collaborative design of clinical cases, created by medical students through a computer-based platform, supports integrating what has been learned and generating knowledge consumed by other students.

The activity diagram of Figure 1 depicts our approach to case-based learning in clinical undergraduate programs. According to this diagram, the approach comprises the following activities:

Consequently, learning occurs in two different scenarios:

In this context, we designed and developed a workflow ecosystem based on the collaborative cocreation of clinical cases, supported by technological platforms and global standards that allow for the standardization, export, and distribution of clinical cases for open and democratic consumption. This ecosystem, which is illustrated in Figure 2, integrates the following platforms:

For the collaborative design of a clinical case, Mosaico provides standardized templates with all the information required. Students would write down relevant information about the clinical case, including age, sex, weight, height, physical exam, vital signs, laboratory tests, images, or videos (see Figure 3).

The web platform allows clinical cases to be generated, guided, and monitored; however, it stores them in a relational database. This does not allow records to be extracted and shared across heterogeneous technologies. In addition, Mosaico’s lack of interoperability features limits the platform’s use by forcing all institutions to use the same technology or make ad hoc integrations if they wish to benefit from the portfolio of clinical cases created.

To provide Mosaico with interoperability capabilities, we proposed including a module that allows the transformation and export of clinical cases into a universal format—one that any other platform can easily read.

This module incorporates an interoperability layer designed and implemented through REST application programming interfaces [18] with HL7 FHIR [19] that offers clinical knowledge representation based on international standards. For this purpose, and according to the case design form provided by Mosaico (Figure 3), a set of HL7 FHIR resources would contain all the standardized clinical case information, making it easily exportable.

Clavy is a learning object repository management platform based on formal grammars. Each Clavy repository comprises (1) a set of object schemata, which describes the structures of the different objects using formal grammars, and (2) the learning objects themselves. In addition, the platform fully supports Extract-Transform-Load (ETL) workflows with user-defined plug-ins and advanced editing facilities.

The tools integrated by Clavy can be accessed on the web and provide user-friendly interfaces to manage the repositories. Using these tools, domain experts without programming knowledge can use different transformations to reorganize the repository’s structure, remove irrelevant information, and add new objects to this repository. For example, Figure 4 shows a screenshot that displays part of a specialized repository in the clinical domain, making it appear as a part of a schema-guided navigation tree, a part of the learning object collection, and a fragment of a particular learning object.

We applied our workflow ecosystem to generate standardized e-learning content for a relevant clinical case. The clinical case to show the proposed approach was related to a patient who consults for chest pain. The case was designed while considering all the clinical backgrounds necessary for its representation through a simulated patient. Furthermore, it was important to consider all the antecedents since the case will later be packaged and distributed to an LMS environment so that other students can consult it as study material.

Fourth-year medical students, along with an academic instructor from the University of Talca in Chile, developed the case in Mosaico by considering the following scenario:

A 58-year-old male patient with a history of arterial hypertension (treated with 20 mg of enalapril every 12 h, which he takes irregularly), who works as an engineer 8 hours per day in an office, consults for oppressive chest pain in the precordial region, which is radiating to the neck and jaw, appearing at rest, and accompanied by coldness and cutaneous pallor that accentuated approximately 8 hours ago. He went to the hospital, was evaluated, and presented the following vital signs:

Moreover, the following complementary tests were incorporated: laboratory exams, chest x-rays, and an electrocardiogram taken in the hospital.

The case was created in Mosaico (see Figure 3) and stored in a MYSQL relational database; all data were then exported as HL7 FHIR resources. Figure 5 shows the resources involved, using a DiagnosticReport resource as the backbone of the case.

Clavy was used to create an ETL pipeline to extract and transform the information created in Mosaico; this information could then be readily exported into IMS-CPs that could be deployed in an LMS, such as Moodle or Blackboard. In this way:

Once the information has been transformed into a structure that suits the instructors’ needs, a suitable Clavy export plug-in can be applied to the resulting repository to package the clinical cases into standardized learning packages that can be loaded into conventional LMSs. In this experience, the clinical case was exported as IMS-CPs and incorporated into Moodle (see Figure 7A). Figure 7B shows an IMS-CP manifest file segment associated with the original Mosaico clinical case.

In this paper, we identified four key elements to enable the effective use of clinical cases in e-learning environments: (1) the effective authoring of cases, (2) the interoperability of authoring tools, (3) the transformation of cases into standard e-learning content packages, and (4) the publication of content packages on widely used e-learning platforms. For this purpose, we proposed a workflow that consistently addresses each of these critical aspects by forming a technology ecosystem supported by (1) the Mosaico collaborative clinical case authoring tool, which supports students and teachers in health care careers with the collaborative design of clinical cases; (2) the HL7 FHIR clinical interoperability standard, which makes it possible to expose relevant information from the clinical cases to third-party tools; (3) the Clavy learning object repository management tool, which allows the transformation of these clinical cases into standardized educational formats; and (4) the Moodle LMS, which makes it possible to deploy the resulting contents for their consumption.

The collaborative design supported by Mosaico [7] allows medical students to create clinical cases through standardized templates with all the information required and the possibility to integrate relevant knowledge from multiple sources. The case study regarding chest pain described in the previous section, in which Mosaico was successfully used to produce a complex clinical case by a group of advanced students under the supervision of a medical instructor, highlights the suitability of Mosaico in orchestrating this type of collaborative clinical case authoring. At this stage, it is essential to highlight Mosaico’s ability to allow the creation of clinical cases by groups of students, which can be shared and exchanged democratically within the academic community. An interesting question to discuss is the wide variety of medical scenarios that can be considered during the design of clinical cases. Since the clinical cases are created by students in the company of an academic instructor, they are never repeated. Each design session has a new case history being shared, regardless of the diagnosis.

Concerning interoperability issues, HL7 FHIR specifications or notations [20] standardize medical records by incorporating an interoperable format that increases the exchange potential of any type of health data source. As exemplified by the case study in the previous section, the interoperability layer added to Mosaico highlights the feasibility of representing complex educational and clinical cases in terms of the HL7 FHIR information models. In this regard, clinical cases created using tools such as Mosaico, represented internally in proprietary and tool-dependent formats, can be offered to third-party tools in standardized and open formats. As shown in the case study, this aspect is essential to facilitating the reuse of clinical cases in many contexts that are not necessarily anticipated. In the work described in this paper, FHIR has allowed the generation of information packages used to build digital medical repositories, which can be deployed in subsequent stages of the design process. A similar environment to the one proposed and aimed at collecting clinical case information based on FHIR notations can be found in the Public Health Automated Case Event Reporting platform [21], which uses laboratory results to extend the description of clinical cases.

Although the availability of clinical cases represented in an interoperable format, such as HL7 FHIR, facilitates their consumption by external tools, a substantial gap prevents their publication in e-learning environments. Indeed, as evidenced in the case study presented, at the syntactic level, the HL7 FHIR information packages produced by Mosaico consist of JSON files and additional associated resources (eg, clinical images). Meanwhile, they represent instances of the FHIR information models at the semantic level. In contrast, information publishable on e-learning platforms follows well-established e-learning standards (eg, IMS-CP or SCORM). Therefore, this is a domain transformation problem at the syntactic level (eg, the transformation of JSON documents into IMS-CP XML manifests) and at the semantic level (eg, the change of FHIR entity graphs into IMS-CP or SCORM organizations). As the case study shows, the Clavy platform demonstrated its ability to undertake the required transformation tasks. Indeed, in this paper, Clavy was able to convert the HL7 FHIR resources produced by Mosaico through automatic procedures, taking the resources as input and generating digital repositories to be edited and processed. In addition, during this transformation stage, the remarkable functionality of Clavy allowed text elements to be converted from Mosaico-based clinical cases into standard terms such as SNOMED Clinical Terms or Logical Observation Identifiers Names and Codes (LOINC), which adds value to the original case descriptions. The advantages of using semantic standardization terminologies have been reviewed in matters such as structured reporting [22] or health care process mining [23].

The final stage consists of generating standard educational specifications, which can be useful when spreading instructional content across diverse health care e-learning environments [24]. In this case, Clavy allowed the export of the clinical cases collected from Mosaico using standard formats such as IMS-CP or SCORM. Standard specifications have played a crucial role in applying e-learning technologies for medical education. Most of these specifications focus on the content sharing of health care learning resources that can be designed and developed collaboratively, as seen in the mEducator initiative [25]. The work presented in this paper is similar to this initiative, relying on learning objects and IMS-CP educational packages to deliver instructional clinical cases to medical students and health care practitioners. In addition, the role of technical standards in medical instructional settings has been reviewed as a part of recent digital innovations introduced in health care education and training [26]. These innovations boost the use of SCORM packages that allow e-learning developers to add enhanced interactive features and assessment procedures to the massive amount of medical knowledge available, which are traditionally stored as plain content items [27]. Moreover, they introduce the possibility of tracking user interactions with these content items to supervise their learning in clinical scenarios [28].

In this paper, we proposed and implemented a workflow to deploy standard e-learning content, combining Mosaico, a tool for collaborative content authoring; HL7 FHIR, a clinical information interoperability standard; and Clavy, a platform for the flexible management of learning object repositories. Through a substantial case study, we demonstrated (1) how clinical cases collaboratively developed in Mosaico can be coded and exported in HL7 FHIR, (2) how this facilitates the reuse of these clinical cases with other tools, and (3) how cases exported in HL7 FHIR can be transformed into IMS-CP e-learning packages using a tool such as Clavy. As a result, we demonstrated the feasibility of deploying collaboratively created content using proprietary authoring tools in conventional open e-learning environments.

Having demonstrated the approach’s feasibility, we intend to apply it to other scenarios of collaborative production of medical e-learning content. Likewise, we want to better integrate the components that constitute the proposed workflow ecosystem to the end users transparently, thus increasing usability and the overall user experience.