Medical science data refer to data generated through basic and applied research, experimental development, as well as raw and derived data obtained through observation, monitoring, inspection, and testing [1]. These data include electronic health records, long-term treatment outcome data, experimental drug performance data, and wearable device data. Medical science data are considered a strategic resource, offering significant scientific benefits such as reducing research costs, accelerating the research process, improving research quality, and increasing the impact of data [2]. For example, electronic health records can not only help doctors make diagnostic and treatment decisions more accurately and quickly, but also assist patients in understanding their own health status and promoting self-management.

According to investigations, approximately 469 and 750 medical science data platforms have been registered on FAIRsharing and Re3data, respectively [3,4]. Despite such widespread availability, user survey reports indicate that the usage and downloads of many platforms are far lower than expected. This gap stems from users’ insufficient awareness of the research field and potential benefits. For example, clinicians still rely heavily on experience and traditional methods for clinical decision-making rather than using data-driven analytical tools. The public perceives medical data as overly specialized, thus overlooking its importance. Platforms often fail to demonstrate the value of data in an understandable and user-friendly way, making it difficult for users to grasp its practical applications.

Existing open health care platforms have challenges such as limited open sharing of data, low standardization, poor data quality, insufficient value mining, insufficient privacy protection, and low usage. The main reason behind these challenges is the inability of both platforms and users to objectively quantify the potential value of data. Platforms suggest identifying the key factors in the formation of data value and allocating resources effectively, while users are unable to carry out data activities based on the true value of data.

In the past, scholars have assessed the value of data in different scenarios, including the assessment of enterprise data value in business contexts [5,6], the evaluation of government data [7,8], and social media data [9] in nontransactional scenarios. Given that platform data are more readily accessible, research has increasingly focused on assessing the value of data from platforms. In terms of assessment theories and perspectives, assessment is usually conducted from the accounting perspective, stakeholder perspective [10], data life cycle or data value chain perspective [11], data traceability perspective [12], and seldom adopts a comprehensive perspective or innovative theories. In terms of evaluation content and methods, the valuation of intangible assets [13,14] typically uses approaches such as the cost method, market method, and income method, with a focus on data value components such as cost and income. In recent years, scholars have introduced revenue monetization, impact-based approaches, experimental and survey techniques, comprehensive multiattribute evaluation, and intelligent methods, which emphasize evaluation from multiple value dimensions and factors affecting data value [15-17].

Medical science data are rich and sensitive enough to generate valuable scientific insights, but the assessment of their value remains fragmented. Initially, Wang and Strong [18] proposed one of the most widely used frameworks for assessing data quality, but due to the complexity of the data quality concept, there is still no consensus on its measurement. Given that different types of data apply to different contexts, they cannot be assessed with a one-size-fits-all approach [19]. Scholars have made significant efforts to assess the value of research data. For example, Schmidt et al [20] developed a framework for assessing the quality of research data. Uribe et al [21] assessed the quality of research data based on the FAIR (Findable, Accessible, Interoperable, and Reusable) principle. Regarding the quality of health care data or information, Feder [22] summarized the dimensions and methods for assessing the quality of EHR data. Kim et al [23] proposed a conceptual model for the quality of health care data. Zhang and Trace [24] explored the quality of self-tracking data. Lasim et al [25] identified accuracy, timeliness, completeness, and consistency as key dimensions for assessing the quality of health care data. Wu et al [26] described the quality of health data in terms of dataset quality, descriptive data quality, and metadata quality. This study assesses the value of medical science data on open platforms based on an integrated perspective and using a combination of qualitative and quantitative methods, aiming to fill the gaps in existing research.

This study aims to construct the index system and model of medical science data value assessment from the perspective of the platform and propose the optimization process of data value assessment and data value enhancement strategy. Specifically, the medical science data open platform includes a medical science data resource base and a knowledge base. The platform perspective refers to analyzing and understanding the role, value, and function of the platform in a specific ecosystem from the perspectives of the platform’s overall architecture, functional design, and interaction with users, data, and other stakeholders.

The questions to be addressed are as follows: (1) How to construct the medical science data value assessment index system and assessment model? (2) How much do different assessment dimensions and indicators contribute to the value of medical science data? (3) How scientific and usable are the constructed assessment models?

The key steps in this process include developing an assessment index system, calculating weights, and conducting comprehensive analyses, which are as follows:

Figure 2 shows the steps to build the evaluation index system.

This study did not require formal ethics committee review, as it exclusively analyzed publicly accessible platform interfaces, technical documentation, and aggregated system-level characteristics, without the collection of identifiable human data. This exemption aligns with the National Health Commission of China Regulations, Article 24, which states: “Ethics review may be waived for studies using publicly available data or records where individuals cannot be identified, and the research does not interact with human participants.” It also aligns with the JMIR policy on nonhuman participants research, which states: “Ethics approval is not required for studies evaluating technology systems, public documentation, or organizational practices where human participants are neither involved nor identifiable.”

To determine the optimal and worst solutions, the positive and negative ideal solution distances and relative proximity were calculated, as shown in Table 5. From Table 5, it can be seen that the National Population Health Sciences Data Center (NCMI; Institute of Medical Information, Chinese Academy of Medical Sciences) has the highest data value, which is close to the ideal level and serves as a valuable benchmark for other platforms.

The weighted formula was used to calculate the scores and composite scores for the different dimensions of data value for each platform:

Where i denotes the ith platform, j the jth dimension, and k the kth indication. Score (Pᵢⱼ) reflects the data value score of the jth dimension of the ith platform, as well as the platform’s overall data value score. Figures 5 and 6 show the differences between each platform’s data value by dimension and their comprehensive scores.

As shown in Figures 5 and 6, there are large differences in data value across the dimensions of intrinsic attribute quality, extrinsic attribute quality, platform service quality, and perceived usefulness. The average score of the platforms is 37.579, with NCMI, MGnify (European Bioinformatics Institute), EBRAINS (Human Brain Project), GSA for Human (National Genomics Data Center), and OpenNeuro (Stanford University) scoring higher than the average, reflecting good overall performance.

The evaluation reveals that (1) Chinese platforms demonstrate data quality strengths but require improvements (eg, NCMI shows slow updates and low availability of machine-readable format; GSA lacks completeness in data attributes/types), (2) platform functionality favors NCMI for comprehensiveness, while OpenNeuro leads in responsiveness, (3) service quality analysis shows NCMI’s advantages in personalization and search capability; however, all platforms need enhancements in service assurance and confidentiality, (4) GSA excels in content novelty/relevance despite overall low data utility across platforms, and (5) GiagDB demonstrates strong usability despite partial FAIR compliance. The model successfully identifies critical gaps in China’s platform data management and proposes targeted optimization pathways, thereby validating its scientific rigor.

This study establishes a novel medical data valuation framework through a 3D (data-platform-user) analysis, integrating service privacy metrics. By combining qualitative and quantitative methods, the hybrid approach reduces subjectivity bias, generating an evaluation model that enables data value quantification and personalized management strategies.

Comparing the results of government data value assessment, the following conclusions can be drawn: (1) indicators related to the data itself, such as the number of datasets, data timeliness, and machine readability, as well as indicators such as system responsiveness and service interactivity, are considered very important, (2) compared with government data, medical science data is private and scientific, so the weight of indicators such as service security, service confidentiality, accessibility, and reusability are more prominent, and (3) in the process of data value realization, the data itself have potential value, but their actual realization requires further development, so indicators related to platform quality and platform service quality, such as search comprehensiveness and service personalization, are also crucial.

This study acknowledges three principal methodological constraints that warrant consideration when interpreting the findings:

At the theoretical level, this study broadens the perspectives, objects, and methods of data value assessment research and refines the strategies for value enhancement of medical science data on open platforms. At the practical level, the assessment model developed in this study can not only help platforms reasonably quantify the potential value of data, identify key factors in the formation of data value, discover deficiencies in data collection and organization, formulate data investment decisions and openness plans, and promote the effective allocation of data resources; it can also help users recognize the potential benefits of the data on the platforms and better use the data resources in their data activities, and help countries to understand the benefits of open medical science data to optimize data regulations and systems.

This study focuses on assessing the value of medical science data from the perspective of platforms, and the results of the study include (1) a data value assessment index system with 5 dimensions and 35 indicators was created by using ISSM, TAM, and the theory of perceived value of consumers, which is scientific, (2) 10 open platforms were selected, data were collected through Octopus and empirically examined using the entropy weight Technique for Order Preference by Similarity to Ideal Solution method, and the assessment model was established and scientifically verified, and (3) the practical application of the proposed model can not only optimize the data value assessment process, but also enhance the data value of the platform from multiple dimensions.