Text

Introduction: The assessment of publications in evidence-based medicine according to risk of bias criteria is challenging, especially in the trending field of complementary and alternative medicine (CAM) [1]. This is why we created the CAMIH database, which is based on Semantic MediaWiki (SMW), as a central hub that allows clinicians to create, discuss and synthesize evidence in CAM. While the multitude of customizability options for content creation is a key advantage of SMW, it also imposes a fixed inherent query system on its users for the synthesis of this content, which is the target application of CAMIH. It requires them to know the syntax of SMW, the case-sensitive names of all properties they intend to query, and the available output format options to structure their results the way they want. This is not feasible, even for experienced users.

In this work, we implement the “CAMIH helper” (https://camih.med.uni-jena.de/helper/) tool that significantly simplifies this task and requires only domain knowledge when generating queries. We highlight the advantages of this new tool when compared with the faceted and semantic search directly provided by SMW.

Methods: CAMIH helper was implemented using the JavaScript framework Vue3 (https://vuejs.org/) in conjunction with Python's Django backend. Design-wise, we aimed for intuitive usability and the generation of only syntactically correct queries, reflected in the following features:

• autocompletion for all fields,
• selection of multiple values for a property automatically applies logical OR
• easy application of the wildcard syntax via checkbox,
• default return message “No results found.” instead of nothing, when no items match the query,
• variable number of conditions and outputs,
• collapsible tutorial.

In SMW, an ontology is established by subobjects. In our case, Arms, Outcomes, and the associated Publication are subordinate to their respective study. Querying properties of these subobjects or returning query results on one of these levels of granularity requires a specific syntax. This functionality is not present in SMW's faceted or semantic search features, so we implemented it in CAMIH helper.

Results: In exploratory tests, we found that users who were unfamiliar with the SMW syntax but versed in evidence-based medicine were able to quickly create valid queries. This included, in particular, finding relevant properties via autocompletion and the intuitive application of logical operators and wildcards.

Discussion: As stated, SMW's query system has shortcomings in the usability aspect. While the inherent features provide some assistance, they cannot be used to design queries directly on wiki pages and they do not support subobject ontologies. However, we found that one advantage of the faceted search feature is that after selecting, e.g., a certain property, you can see how many results your query will yield if you choose to narrow it down further. Since our helper database is not directly linked to the CAMIH database, we cannot provide this feature right now.

Conclusion: In this work, we presented the web-based CAMIH helper tool, which is designed to make the SMW query system more accessible. Our preliminary usability tests showed that domain experts, who have never worked with SMW, were able to quickly construct valid queries.

The authors declare that they have no competing interests.

The authors declare that an ethics committee vote is not required.

Literatur