EHDS Hospital Readiness Survey

Section 01

Introduction

About the survey · Respondent profile · n = 62 hospitals

In early 2026, the Belgian Health Data Agency (HDA) conducted a readiness survey to assess how prepared Belgian hospitals are for the European Health Data Space (EHDS), the EU regulation that is designed to standardize, secure, and facilitate the exchange of health data across Member States.

The survey was structured in two parts. Part 1 asked CMIOs and data architects to rate their hospital's maturity across 10 Data Management & Governance themes on a scale of 1 (no structure) to 4 (fully mature/advanced). Part 2 assessed the presence of 6 technical data architecture components using a binary yes/no format.

Supplementary open-ended questions captured the main challenges and planned improvements at each hospital, later coded into 18 challenge buckets and 11 improvement buckets for quantitative analysis.

Where relevant, differences between hospital groups were tested for statistical significance. Statistically significant findings are highlighted throughout the presentation.

Based on the survey findings, the HDA has developed a structured action plan to support hospitals in meeting EHDS requirements by 2029. Each recommendation comes with concrete actions, an effort estimate, and guidance on who in the hospital should lead it.

62

Belgian hospitals surveyed

Two-part survey sent to CMIOs & data architects
62 of 159 Belgian hospitals responded (39%)
Results as of 27 April 2026

The 62 hospitals form a representative sample of the Belgian landscape in terms of size, regional distribution, status (private/public) and type (university/non-university).

n = 159

Goals of this report

1

Support hospitals in defining pragmatic improvement roadmaps

2

Support policymakers and ecosystem actors in targeting supporting initiatives

3

Support alignment between hospital strategies and national and European EHDS implementation efforts

4

Baselining and benchmarking between hospitals

EHDS Familiarity

How familiar are Belgian hospitals with the EHDS regulation?

The approach

1

Targeted EHDS readiness survey

A structured two-part survey completed by CMIOs and data architects across Belgian hospitals.

❯

2

EHDS survey results deep-dive

Hospitals willing to collaborate are invited for an on-site deep dive to explore findings in detail.

❯

3

Results & recommendations

Actionable guidance including priority gaps, maturity benchmarks, and pragmatic next steps.

Section 1: General Data Management & Governance

Data Interoperability
Data Governance & Responsibilities
Data Quality Management
Metadata Management & Documentation
Data Exchange & Access for Secondary Use
Data Privacy & Legal Compliance
Data Security & Infrastructure
Data Integration for Research & Analytics
Data Literacy & Training
Transparency & Patient Engagement

Multiple choice questions: 4 maturity levels
Open questions: Challenges & improvements

Section 2: Data Architecture Maturity

Central (Standardized) Data Repository
Mitigation of Small Cells and Encoding Service of eHealth (Pseudonymization)
Metadata Management Tool
Data Exchange
Data Quality Checks
Other Means of Capturing Data

An additional question on Electronic Health Records (EHR) was asked, but only via open questions (what works well & what doesn't), not as a Yes/No question on presence at the hospital in question, like for the other 6 components.

Multiple choice questions: Component used Yes/No
Open questions: What works well & what doesn't work well

EHDS survey results deep-dive

Structured interviews

Follow-up conversations with hospitals that volunteer to participate

Score deep-dive

In-depth review of individual maturity scores and the reasoning behind them

Quick wins

Concrete priority improvements specific to each hospital

Roadmap alignment

Mapping existing initiatives to the EHDS compliance timeline

Participation is voluntary · sessions are confidential

Results & recommendations

The results and recommendations are presented in this interactive HTML report. Navigate using the menu at the top or explore each section below.

Introduction
survey design, respondent profile, EHDS familiarity
Executive Summary
key findings at a glance, where Belgium stands
Maturity Overview
data governance maturity vs technical maturity scatter, broken down by hospital profile
Data Governance Maturity
radar chart with averages across all 10 themes, broken down by hospital profile
Data Architecture Maturity
component adoption, broken down by hospital profile
Challenges
18 challenge buckets ranked by prevalence
Improvements
11 planned improvement buckets
By Theme
challenges and improvements mapped to governance themes
Recommendations
sunshine roadmap and detailed recommendation cards with actions and guidance

All data is based on n=62 Belgian hospital responses collected in early 2026.

Do you want to learn more about metadata, data quality or the EHDS regulation? Visit the HDA Academy to sharpen your knowledge.

Section 02

Executive Summary

Where Belgium's hospitals stand against the 2027 to 2031 EHDS timeline

Block 01

The overall picture

Partial maturity, on a clock that started running on 26 March 2025.

Belgian hospitals sit at partial maturity on both axes the survey measures. The average Data Governance score across all 10 themes is 2.39 out of 4, which sits between "emerging" and "structured", and average technical component adoption is 0.49 out of 1. Most hospitals cluster between these two levels, and neither group reaches the level of integration the EHDS assumes in its operational provisions.

EHDS compliance timeline

Mar 2025

Entry into force

Aug 2026

Today

Mar 2027

General date of application; deadline for key implementing acts

Mar 2029

Rules on secondary use will also start to apply for most data categories (e.g. EHR data)

Mar 2031

Rules on secondary use will also start to apply for the remaining data categories (e.g. genomic data).

Mar 2035

Third countries can join HealthData@EU for secondary use

Data Governance Maturity

0.00/ 4

Average across all 10 themes

Technical Component Adoption

0.00/ 1

Average of 6 architecture components

Block 02

Strong themes reflect what regulation and national programs have already required

Where Belgian hospitals score highest, an external regulatory driver was already in place.

The three themes scoring highest are all areas where hospitals have had a binding external driver for some time. These results are largely as expected. They confirm that Belgian hospitals respond to regulation when it is specific and enforceable, and that the foundations for safe secondary use processing (confidentiality, security, structured exchange formats) are broadly in place.

2.9/4

Data Privacy & Legal Compliance

→

GDPR Binding since 2018

2.9/4

Data Security & Infrastructure

→

NIS2 Transposition since 2024

2.8/4

Data Interoperability

→

HL7 / FHIR National programs ongoing

Block 03

EHDS turns previously soft areas into legal obligations

The themes scoring lowest are precisely the ones the regulation now mandates.

The themes scoring lowest are precisely the ones the EHDS converts from best practice into compliance requirements. The gap between the strong and the weak themes in this survey is not coincidental: it tracks the boundary between obligations that have existed for years and obligations that are new.

1.8/4

Metadata Management & Documentation

→

Art. 60 Dataset description must be communicated to the HDAB and verified at least annually.

2.1/4

Data Quality Management

→

Art. 78 Data quality & utility label is mandatory for datasets collected with public funding

27%

Pseudonymisation / metadata tooling in place

→

Art. 66 Low uptake may constrain hospitals' ability to supply pseudonymised data for secondary use

Transversal enabler

Data Literacy & Training: lever for every obligation above

1.9/4

Second lowest theme in the survey

No direct EHDS obligation, but a precondition for all the above. Accurate dataset descriptions (Art. 60), quality labels (Art. 78) and EHDS transition capacity all require staff who understand the data they hold. Today, training reaches IT and research but rarely clinical or operational staff.

Technical signal

Other Means of Capturing Data: widespread adoption masking infrastructure gaps

76%

Highest of all 6 components

76% of hospitals indicate using other means to capture data, the highest adoption rate of all six components, yet this reflects workarounds (CSV extractions, ad-hoc scripts), not standardised infrastructure. Alongside 16% for Metadata Management and 15% for Pseudonymisation, hospitals can move data when pressed but not in the structured, documented form Art. 60 requires.

Block 04

A polarized sector, not a uniform baseline

The sector is splitting along size and region, rather than converging on a shared starting line.

Walloon hospitals are proportionally twice as likely as Flemish ones to indicate low data governance maturity as well as low technical component adoption. The EHDS does not differentiate Article 60 obligations by hospital size, type or region. A meaningful share of Belgian hospitals will therefore need to close a larger gap, on the same timeline, than the sector average suggests.

Technical Component Adoption

Data Governance Maturity

low high low high

18%

42%

24%

16%

Where differences are statistically significant

(click on the tiles to see which themes show significant differences by category)

Governance Maturity

Size✓

Region✓

Type✗

Status✗

Significant on 3 themes. Larger hospitals score higher on Data Governance & Responsibilities, Data Exchange & Access for Secondary Use, and Data Security & Infrastructure.

Significant on 1 theme. Flemish hospitals score higher than Walloon hospitals on Data Integration for Research & Analytics.

No significant difference on any of the 10 governance themes. University hospitals trend higher across themes, but not significantly.

No significant difference on any of the 10 governance themes between public and private hospitals.

Technical Architecture

Size✗

Region✗

Type✓

Status✗

Significant on 4 components. University hospitals show higher adoption of the Metadata Management Tool, Data Quality Checks, Data Exchange, and Other Means of Capturing Data.

No significant difference on any of the 6 architecture components by hospital size.

No significant difference on any of the 6 architecture components by region.

No significant difference on any of the 6 architecture components between public and private hospitals.

Block 05

Hospitals' own diagnosis vs. their planned response

The top cited challenge is structural, but the planned investments are predominantly technical.

When asked what stands in the way, 73% of responding hospitals cite lack of time, incentives or recognition as the single most prevalent challenge across all 18 coded categories, yet none of the most-cited planned improvements address it directly. Instead, 81% plan to invest in data quality management frameworks and 76% in interoperability standards, both technical fixes for what is, at its root, an organisational problem. Closing the gap to EHDS by 2029 will likely require addressing both, the structural and the technical.

Top cited challenges

↔

Top planned improvements

71%

Incomplete adoption of interoperability standards

76%

Adoption of interoperability standards (FHIR, SNOMED CT, LOINC, OMOP)

60%

Poor data quality at the source

81%

Data quality management frameworks

47%

Legacy systems and technical debt

Spread across

55%Structured and coded data capture at the source

45%Central data platform for analytics & secondary use

73%

Lack of time, incentives, or recognition

Partial map

60%

Data literacy & training (addresses recognition, not time or incentives)

Block 06

From diagnosis to action

63 prioritized recommendations across 5 strategic goals and 4 phases. Every recommendation comes with concrete actions and implementation guidance.

In response to these findings, the HDA has developed a four-year action plan of 63 prioritized recommendations, structured across five strategic goals and four implementation phases (2026-2029). Each recommendation comes with concrete actions, implementation guidance, an effort estimate, a priority level, and an indication of who in the hospital should lead it.

Phased over four years

2026

Foundation & initial implementation

29

recommendations

2027

Expansion & integration

15

recommendations

2028

Optimization & compliance

16

recommendations

2029

Full deployment & continuous improvement

3

recommendations

Structured across five strategic goals

1

Set the ambitions and vision for data and obtain leadership sponsorship

5

2

Define who does what when it comes to Data and AI in the organization and implement this governance model. Additionally reinforce and promote Data and AI literacy.

17

3

Redefine a data offering that aligns with the data vision and strategy

6

4

Realize qualitative and efficient internal and external data sharing, supported by data (quality) management

17

5

Support the exchange of data and creation of data products by implementing the right tooling and infrastructure.

18

Every recommendation includes

Concrete actions

Specific steps to execute

Attention points

Pitfalls and implementation guidance

Importance level

Crucial, important, or valuable; assigned based on how directly the recommendation maps to EHDS legal requirements

Effort estimate

Low, medium, or high in person-days

Implementation level

To whom in the hospital the recommendation applies (e.g. executive committee, management, staff, IT, legal)

Timing

Target year aligned to EHDS milestones

→ View all recommendations and roadmap

Section 03

Maturity Overview

Data Governance Maturity (1–4) vs Technical Maturity (0–1) · n = 62 Belgian hospitals

Overview

By Size

By Status

By Region

By Type

All hospitals · bubble size ∝ overlap count · dashed lines = averages (DG 2.39, Tech 0.49)

n = 62

Coloured by hospital size · pie slices show composition where points overlap

n = 62

Coloured by hospital status (Private / Public)

n = 62

Coloured by region (Flanders / Wallonia / Brussels)

n = 62

Coloured by hospital type (Non-University / University)

n = 62

r = 0.39

Moderate positive correlation between DG and Technical maturity

Hospitals strong on governance tend to be strong technically too (Pearson r = 0.39, p < 0.001).

Wallonia disproportionately low

7 of Wallonia’s 19 hospitals fall in the Low DG + Low Tech quadrant, compared to only 6 of Flanders’ 36 — proportionally twice as likely (37% vs 17%)

26 · 15

High-high vs Low-low quadrant (split at DG mean 2.39, Tech mean 0.49)

11 Low-DG / High-Tech · 10 High-DG / Low-Tech — polarisation rather than a uniform baseline

Section 04

Data Governance Maturity

Part 1 — Data Management & Governance · 10 themes · scale 1–4

Radar

By Size

By Region

By Status

By Type

Min / Average / Max maturity per theme · all 62 hospitals · scale 1–4

n = 62

1 = Initial
2 = Emerging
3 = Structured
4 = Integrated

Mean maturity score per size group · S = 8 hospitals · M = 30 · L = 24 · blue points indicate statistically significant differences

n = 62

Hospital size makes a confirmed difference on three of the ten governance themes. On Data Governance & Responsibilities, large hospitals score higher than small hospitals, and medium hospitals also score higher than small hospitals. On Data Exchange & Access for Secondary Use and on Data Security & Infrastructure, large hospitals score higher than medium hospitals. (Kruskal-Wallis: Data Governance & Responsibilities p = 0.011, Data Exchange p = 0.041, Data Security p = 0.020; Dunn post-hoc with Bonferroni correction: Large > Small p = 0.009, Medium > Small p = 0.034 for Data Governance; Large > Medium p = 0.041 for Data Exchange; Large > Medium p = 0.032 for Data Security.)

Mean maturity score per region · Brussels = 7 · Flanders = 36 · Wallonia = 19 · blue points indicate statistically significant differences

n = 62

Flemish hospitals score higher than Wallonian hospitals on Data Integration for Research & Analytics — the only governance theme where a regional difference is confirmed. No other regional difference is significant across the remaining themes. (Kruskal-Wallis p = 0.005; Dunn post-hoc: Flanders > Wallonia, p = 0.006.)

Mean maturity score per status · Private = 48 hospitals · Public = 14 hospitals

n = 62

Whether a hospital is private or public makes no meaningful difference on any of the 10 governance themes — both groups perform similarly across the board. This means ownership model is not a driver of governance readiness; factors such as size and complexity matter more. (Kruskal-Wallis, all p > 0.05.)

Mean maturity score per hospital type · Non-University = 43 · University = 19

n = 62

University hospitals tend to score higher than non-university hospitals on most governance themes, but this difference is not large enough to be statistically confirmed. The difference is most visible on Data Integration for Research & Analytics (university mean 2.63 vs non-university 2.40), which is consistent with university hospitals having stronger research infrastructure. This trend is directional and worth monitoring even if not yet statistically confirmed. (Kruskal-Wallis, all p > 0.05.)

2.39

Overall average DG Maturity across all hospitals and themes

Most hospitals cluster between levels 2 and 3 — partial/emerging maturity

Privacy ▲

Data Privacy & Legal Compliance and Security are the strongest themes (median 3.0)

Likely GDPR-driven · no significant difference by status or region

Size ▲

Size is the strongest grouping variable: significant on 3 of 10 themes

Status (Private/Public) shows no significant difference on any theme

Section 05

Data Architecture Maturity

Part 2 — Adoption of 6 binary architecture components · n = 62

Overview (all 62)

By Size

By Region

By Status

By Type

All 62 hospitals · % with component present · sorted by adoption rate · Hospital Type is the strongest significant grouping variable (Fisher exact)

n = 62

Adoption rate (%) per size group · S = 8 hospitals · M = 30 · L = 24

n = 62

Hospital size does not meaningfully predict which architecture components are in place — the differences observed between size groups are within the range of chance. However there are notable numerical gaps that do not yet reach significance: large hospitals are more than twice as likely to have a Central Data Repository (79%) compared to small hospitals (38%), and are more likely to have a Metadata Management Tool (25% vs 10% for medium hospitals). These gaps may become significant as the sample grows. (Fisher exact, all p > 0.05.)

Adoption rate (%) per region · Brussels = 7 · Flanders = 36 · Wallonia = 19

n = 62

Region does not meaningfully predict architecture component adoption — the numerical patterns visible in the chart are not large enough to be statistically confirmed. Some patterns are visible — Flemish hospitals are more likely to have Data Quality Checks (56%) than Wallonian hospitals (32%), and Brussels hospitals have a notably higher Metadata Management Tool adoption (43%) compared to Wallonia (5%) — but none of these reach statistical significance. (Fisher exact, all p > 0.05.)

Adoption rate (%) per status · Private = 48 hospitals · Public = 14 hospitals

n = 62

Whether a hospital is private or public makes no meaningful difference on any of the 6 architecture components. The most notable numerical gap is for Pseudonymization, where public hospitals show slightly higher adoption (29%) than private hospitals (10%), possibly reflecting greater exposure to secondary use obligations. However this difference is not statistically confirmed. (Fisher exact, all p > 0.05.)

Adoption rate (%) per hospital type · Non-University = 43 · University = 19 · blue points indicate statistically significant differences

n = 62

University hospitals are better equipped than non-university hospitals on 4 of the 6 components — a difference large enough to be statistically confirmed. The gap is most striking for the Metadata Management Tool: 42% of university hospitals have one in place compared to only 3% of non-university hospitals. University hospitals also lead on Data Exchange (95% vs 64%) and Other Means of Capturing Data (95% vs 67%). Hospital type is the only grouping variable that significantly predicts architecture adoption. (Fisher exact, all p < 0.05.)

0.49

Average Technical Maturity (mean of 6 binary components) across all hospitals

Considerable spread across hospitals: standard deviation 0.23, meaning many hospitals have adopted only 1–2 of the 6 components while a small number have adopted all or none.

Type ✓

Hospital Type is the only significant differentiator — 4 of 6 components differ significantly by Type

Components where the university vs. non-university gap is statistically confirmed: Other Means (p = 0.025), Metadata Tool (p = 0.001), Data Exchange API (p = 0.014), Data Quality Checks (p = 0.013)

76%

76% of hospitals use additional data capture tools beyond their core EHR

Widespread adoption signals core systems are failing to capture what hospitals need — driving fragmentation, double entry, and data silos rather than reflecting genuine capability

Section 06

Key Challenges

18 challenge buckets coded from open-text responses · ranked by prevalence (n = 62) · darker bars = top 8

Click on a bar to learn more about each challenge.

n = 62

73%

“Lack of time, incentives, or recognition” is cited by 45 of 62 hospitals

The single most prevalent challenge — structural, not technical

71%

“Incomplete adoption of interoperability standards” ranks second (44 hospitals)

Closely followed by poor data quality at source (60%) and legacy systems (47%)

8%

“Limited DPO capacity” is the least cited challenge (5 of 62 hospitals)

Low prevalence may reflect under-reporting rather than absence of the issue

Section 07

Planned Improvements

11 improvement buckets coded from open-text responses · ranked by prevalence (n = 62) · darker bars = top 7

Click on a bar to learn more about each improvement.

n = 62

81%

“Data quality management frameworks” is the top planned improvement (50 hospitals)

Direct response to data quality ranking as top technical challenge

76%

“Adoption of interoperability standards” planned by 47 hospitals (FHIR, SNOMED CT, OMOP…)

Mirrors interoperability as the second most cited challenge

34%

“Cybersecurity & NIS2 compliance” is the least cited planned improvement (21 hospitals)

Possibly already addressed or handled via separate compliance tracks

Section 08

Challenges & Improvements by Theme

% of hospitals citing any challenge / planning any improvement in each theme · includes primary + secondary theme mappings · n = 62

Challenge prevalence Planned improvement prevalence Hover bars · see contributing buckets per theme

← Challenge %

Theme

Improvement % →

93.5%

Data Interoperability

87.1%

91.9%

Data Integration for Research & Analytics

87.1%

88.7%

Data Quality Management

85.5%

90.3%

Data Exchange & Access for Secondary Use

75.8%

88.7%

Data Governance & Responsibilities

69.4%

80.6%

Data Literacy & Training

77.4%

53.2%

Data Privacy & Legal Compliance

77.4%

40.3%

Metadata Management & Documentation

37.1%

71.0%

Data Security & Infrastructure

33.9%

54.8%

Transparency & Patient Engagement

35.5%

8

In 8 of 10 themes, more hospitals cite a challenge than plan an improvement

The gap is widest for Data Security (69% challenge vs 32% improvement) and Data Exchange (89% vs 76%), meaning recognition of the problem has not yet translated into planned action

39%

Only 39% of hospitals cite Metadata Management as a challenge — the lowest of any theme

Yet Metadata Management is the lowest-scoring governance theme (mean 1.8/4) and a foundational EHDS requirement. Low challenge rate likely reflects low awareness, not genuine readiness

19pp

Data Governance shows the widest gap between challenge recognition and planned action of any governance theme

87% of hospitals cite Data Governance as a challenge, yet only 68% plan improvements — a 19 percentage point gap. Data Literacy & Training shows a similar pattern. Both themes depend on sustained investment in people, roles and culture rather than tooling alone. Without closing these gaps, technical investments in interoperability and secondary use are unlikely to deliver durable results at the hospital level.

Section 09

EHDS Readiness Roadmap

A sunshine roadmap to support hospitals in transitioning from the current level of data maturity to the desired level of maturity — 63 prioritised recommendations across 11 themes.

A sunshine roadmap has been elaborated to support hospitals in gradually transitioning from the current level of data maturity to the level of maturity required under the EHDS. The sunshine roadmap outlines the order of deployment and prioritization of the different recommendations over time, classified by data theme and following different goals that the EHDS tries to achieve. Click on each bubble to see details in terms of concrete actions, implementation guidance, level of implementation, priority and effort.

Concrete Actions

Attention Points

Theme

Priority

Effort

Low = < 5 days · Medium = 5–50 days · High = > 50 days

Level of Implementation