Skip to article

GapBridge Institute

← All articles

GapBridge Institute Articles

How to review field data before it reaches a dashboard

A practical review routine for spotting missing reports, checking unusual values and keeping corrections traceable.

Two participants review a laptop during a workshop.

A polished dashboard can make rough data look more trustworthy than it is. Before reporting, separate two questions: did the records arrive, and are they ready to use? A missing submission, an incomplete record and a confirmed zero tell different stories. Treating all three as zero can make reporting delays look like a drop in activity.

Agree the checks before collection starts

Give field teams one definition for each required field and one reporting deadline. If a field does not apply, let the collector record that explicitly instead of guessing. Give each submission a stable identifier that is not a person’s name. Agree who can correct an entry, who can approve it and where the original source can be checked. These decisions are easier to make before the first reporting cycle than when the deadline is already close.

Triage the records in three passes

First check coverage and timeliness: compare expected submissions with those received and note which are late. Next check consistency: look for duplicate records and compare totals with their components. Finally, check questionable entries against the original source when you have permission to access it. A large change is a reason to look, not a reason to change a value. Ask the collector whether there is an explanation and record what the reviewer confirmed.

For example, imagine a training site reports 38 completed practice assignments while its breakdown by session adds up to 28. The reviewer does not replace 38 with 28. They return the record with a specific question about the total and the session rows. The collector checks the source tally and either corrects the record or explains why the figures use different definitions. The reviewer can then approve it or keep it out of the reporting cut. These figures are fictional.

Keep the review visible in the report

Use separate statuses for submitted, returned and approved records. Include only approved records in reported totals and show how many expected submissions were still missing or under review at the cutoff. Keep a restricted correction log recording what changed, why and who approved it. Share aggregate results with a note about incomplete coverage; do not include raw records or identifiers in a general dashboard. At the next review meeting, use recurring corrections to improve form instructions or validation rules, rather than asking reviewers to fix the same problem every month.

Further reading: WHO Toolkit for Routine Health Information Systems Data: Data Quality Review.