The European Health Data Space creates new requirements for secondary use of health data, including a data quality and utility label for datasets intended for reuse. Data holders, data users and health data access bodies therefore need skills and organisational capacity to assess, document and communicate data quality. A cross-sectional survey published in the Journal of Medical Internet Research gathered 64 responses from 44 institutions across 18 European countries and identified substantial skills, training and organisational gaps.
Health Data Quality Affects Daily Work
Health data forms a routine part of work for most respondents. More than four in five interacted with health data at least weekly, while just over half did so daily. Data quality carried clear operational importance, with a large majority rating it as moderately to absolutely critical for their work. Nearly one in five said decisions could not be made without first ensuring the quality of the data.
Poor data quality also limited effectiveness for most respondents. Missing or inconsistent data formed the most common challenge, alongside delayed availability, weak standardisation and insufficient metadata. Data quality problems were frequent for a substantial share of participants, while a smaller group felt consistently held back by such issues.
Different stakeholder groups experienced these problems in distinct ways. Health data access bodies showed the highest frequency of daily data use and were most likely to indicate that poor data quality severely limited their work. Data users also reported operational effects. Data holders more often viewed data quality issues as manageable, although they represented a varied group with different responsibilities across data capture, curation and management.
Must Read: Radiology Training Faces an AI Skills Shift
Practical Experience Does Not Ensure Competence
Hands-on exposure to data quality tasks was common, but capability remained uneven. Almost eight in ten respondents had previously assessed the quality of a health dataset. Self-assessed understanding most often reached a moderate level, while a smaller group rated their knowledge as high. No respondents selected no understanding.
Experience did not remove persistent gaps. Respondents had encountered errors including missing data, inconsistent formats and poor documentation, but only a minority said they often resolved such problems effectively. Most resolved them only sometimes, while some struggled to resolve them at all.
The most frequently identified skill gaps involved developing and applying data quality metrics, data quality auditing and reporting, metadata management and documentation. Other needs included training in relevant standards, use of data quality improvement software, advanced data analysis, implementation of governance policies, data integration and methods for communicating quality issues.
Organisational readiness also remained limited. Only a small minority reported clearly defined data quality roles. More than two thirds lacked a dedicated data quality manager or team. Many described governance structures as vague or inconsistent, with incomplete datasets, insufficient training and outdated infrastructure adding further barriers.
Training Needs Point to Role-Specific Support
Learning preferences favoured practical and flexible formats. In-person workshops and tutorials were the most favoured options, followed by webinars and self-paced courses. Reading materials were least preferred. Most respondents could dedicate one to two hours per week to training, while none were willing to commit more than five hours.
Participants valued schedule flexibility, continued support and user-friendly training platforms when choosing learning programmes. Tutor access, affordability and certification also mattered, although less strongly. These preferences point towards concise formats that fit alongside professional responsibilities rather than extensive training commitments.
Role-specific differences shaped the training agenda. Data users and health data access bodies tended to rate their knowledge more highly and showed greater confidence in identifying and addressing data quality issues. Data holders showed broader variation and reported more structural challenges. Their responsibilities around data capture and curation place them close to the start of the data life cycle, where problems can affect later reuse.
Priority areas include applied training in data quality metrics and audit frameworks, stronger metadata management and documentation practice and governance support focused on clearer roles. These areas align with the most common gaps and with requirements around data quality and utility labelling under the European Health Data Space.
European health data stakeholders show frequent engagement with health data and substantial practical experience, but important gaps remain in systematic data quality competence. Poor data quality limits effectiveness, while unclear roles, limited governance and lack of dedicated personnel constrain organisational readiness. Training needs are practical, time-sensitive and uneven across stakeholder groups. Data holders appear to need particular support because of their position in data capture and curation. Effective implementation of data quality and utility labelling depends on both individual skills and stronger organisational structures.
Source: Journal of Medical Internet Research
Image Credit: iStock
References:
Declerck J, Eklund N, Sáez C et al. (2026) Health Data Quality Skill Gaps and Training Needs Among European Health Data Stakeholders: Cross-Sectional Survey. J Med Internet Res, 28:e86878.
