Choose the Denominator Before Comparing Safety Data

Name what was counted

A classroom case contains six housekeeping observations from twelve completed review visits in one period. Another period contains eight observations from twenty visits. The definition stayed unchanged, every visit was documented, and one visit could produce more than one observation. These are invented assumptions for this exercise.

The larger second count does not establish that observations were more frequent per visit. Label the numerator recorded observations and the denominator completed review visits. Substituting employees, work hours or calendar days would create a different measure. Write the period and unit beside the calculation.

Use the same scale twice

On a scale of ten visits, these become 5 and 4 observations per ten visits. Multiplying by ten changes presentation, not the underlying comparison. This is a classroom convention, not a regulatory incident-rate formula, an audit score or a recommended organizational target. Use the named units when explaining either result.

Look inside the denominator

The comparison assumes visits are sufficiently comparable for the question. If later visits were shorter, covered fewer areas or used a different checklist, the interpretation would need qualification. More complete reporting could also increase recorded observations without proving that conditions deteriorated.

Now suppose there were no completed visits in the second period. A rate per visit would be undefined because the denominator is zero. Report the absence of visits rather than displaying a zero observation rate. A reassuring zero would conceal the fact that no relevant measurement took place.

Write the claim and its limit

A defensible summary is: “The dataset records 0.50 observations per completed visit in the first period and 0.40 in the second under the stated counting convention.” This does not establish causation, severity, hazard coverage or the quality of the response to any finding.

For each practice comparison, ask whether the numerator and denominator cover the same people, activity and period. Convert presentation scales before comparing. If the underlying definitions differ, explain the mismatch rather than ranking the results. A useful answer may identify what must be clarified before any meaningful calculation can be made.