The Rounds Journal · Health data
Healthcare data governance: who owns the decision behind the dashboard?
Before changing a dashboard, clarify who defines the measure, approves access and checks the data. Governance turns those responsibilities into a repeatable process.
Healthcare data governance is the agreement about how information is defined, accessed, checked and used. A dashboard can be technically correct and still confuse its readers if two teams define the same measure differently or nobody owns a change to the source.
Start with one report that causes disagreement
Imagine operations and finance discussing different counts of completed appointments. One count includes all booked slots; the other includes only attended appointments recorded before a monthly cut-off. Neither disagreement nor a software upgrade tells you which definition should answer the decision.
Write the question first. Then document the event being counted, the inclusion rules, the period and the source. Ask who is responsible for approving that definition and communicating changes.
Assign responsibilities that people can use
| Responsibility | Question to resolve |
|---|---|
| Definition | Who approves what the measure means and when it changes? |
| Collection | Who understands how source records are created? |
| Quality | Who investigates missing, duplicated or inconsistent records? |
| Access | Who approves the minimum information each role needs? |
| Use | Who explains the limitations before a decision is made? |
These are responsibilities to allocate, not mandatory job titles for every organization. One person may hold several responsibilities, but the accountability should still be visible.
Create a short data dictionary
For each field, record its name, meaning, allowed format, unit, source and known limitations. For a measure, add the calculation and denominator. Keep the version and approval date. If “discharge date” changes from a planned date to an actual event date, a trend may change even when the service has not.
The data-cleaning example shows why appointment identifiers and missing-value rules need definitions before calculation.
Use approved access and change processes
A training task does not justify copying workplace records into a personal device or public tool. Use approved access, storage and sharing arrangements. When practice data is enough, use synthetic or properly published public data. Removing names alone should not be assumed to make a dataset safe to share.
Specific privacy, retention and cross-border requirements depend on the information, organization and jurisdiction. Ask the responsible governance or legal team for the applicable rules rather than treating this article as a compliance checklist.
Make changes explainable
- Record the issue and the proposed change.
- Identify affected measures, reports and users.
- Have the appropriate owner approve it.
- Check the effect on a small example before release.
- Tell users what changed and whether comparisons remain meaningful.
A reliable report needs technical work and organizational judgment. If you are drawn to the analytical work, explore health data study; if the challenge is coordinating responsibilities, management development may also be relevant.
Explore study around your background and goals →
Sources and scope
WHO health-service data quality guidance
Sources checked 15 November 2026. Examples and planning exercises are editorial guidance, not reported outcomes or individual admission decisions. Confirm current requirements directly before applying.
Follow a measure from its source to a decision
Imagine an outpatient team sees 120 completed appointments on its monthly report while another department reports 112. Before asking which dashboard is right, trace both back to their source. One report might count appointments marked complete before the monthly cut-off. Another might count records updated later or use the date a service was billed. Ask the report owners for definitions, a publication date, and the person to contact about changes.
Write a measure record that identifies the decision the measure supports, who approves the definition, how the data is extracted and how the result is checked. Use a fictional example for training. If you compare real records in an approved setting, check the authorised purpose and use the smallest amount of information needed to explain the difference. Write the agreed definition and change date next to the dashboard so users can interpret results from both periods.
Agree how changes and corrections will be handled
Suppose a clinic moves to a new booking system and two status values are combined. Ask how the values map before the new records appear in a report. Compare the old and new calculations on a small set of approved examples. Decide whether to adjust earlier results, show a break in the series, or keep separate reports for different purposes. Record who made that decision and how it was communicated.
For a correction to an issued report, note what was wrong, how it affected the results and which version readers should use now. Tell the people whose decisions may depend on it. Agree who can make a correction and which change requires approval. A correction record matters when a figure is quoted outside the dashboard and the original version has already been circulated.
Review the governance around the work, not just the paperwork
In a review meeting, take one recent dashboard update and ask who asked for it, who checked it and who can explain it now. If several teams own different parts, name a contact for resolving differences. Ask a colleague to reproduce one measure from its definition and identify what they would do if an input was missing. Use that walkthrough to clarify ownership, permissions or documentation where the process actually breaks down.