A regional director asks two administrators for their call-answer rates and gets two numbers that look nearly identical. Only later does it become clear that one building counts calls answered by voicemail as answered, and the other does not. The comparison was meaningless, but it looked authoritative.
This happens in multi-facility groups constantly. When metrics share names but not math, dashboards mislead. Fixing it takes some unglamorous work: agreeing on definitions.
Why definitions drift
Buildings evolve their own ways of working. Different systems, different habits, different histories. A report built by one person in one building becomes the standard because it is the one that exists. Over time, small differences in what is included, excluded or timed produce large differences in results.
Start with a short data dictionary
For every metric on your group dashboard, write down:
- Name and a plain-language description.
- Formula, including what is counted in the numerator and denominator.
- Source system and field.
- Inclusions and exclusions, such as business hours only or all hours.
- Time period and how it is calculated.
- Owner, the person who can answer questions.
One page per metric is plenty.
Examples of definitions that need agreement
Census
Is it counted at midnight, at the start of the day, or as an average? Are residents on hospital hold included? How are same-day admissions and discharges handled?
Call answer rate
Does a call answered by voicemail count? Do after-hours calls count? What about calls transferred between lines?
Call-light response time
Is it measured from the time of the call to staff arrival at the room, or to the first acknowledgment on a device? Are calls canceled at the resident's room included?
Worked hours per resident day
Which roles are included? Are paid breaks counted? How are agency hours treated? Do you divide by census or by licensed beds?
Days in A/R
Which accounts are included, and how is the denominator calculated?
None of these has a single correct answer. What matters is that every building uses the same one.
Process for agreeing
- Form a small working group with an operations leader, a finance leader, a clinical leader and one or two administrators.
- Draft definitions for your core metrics, noting where buildings currently differ.
- Test on real data. Calculate each metric both ways for a few buildings and see how much the answer changes.
- Decide and document. Where definitions differ, choose the one that best reflects the purpose of the metric.
- Communicate and train. Make sure everyone knows the definitions and why they were chosen.
Handle local differences
Sometimes a building has a legitimate reason to differ, such as a different phone system that cannot report a certain field. Document the exceptions and, if possible, note them on the dashboard so that readers know.
Automate the math
Manual calculation invites inconsistency. When definitions are built into the data pipeline once, every building's numbers are calculated the same way automatically. This is one of the strongest arguments for centralized analytics: the definition lives in one place and can be changed in one place.
Version your definitions
When a definition changes, record the date and consider restating historical data for comparability. If you cannot restate it, mark the break on trend charts. Otherwise, an apparent improvement may simply reflect a changed formula.
Build trust through transparency
Administrators are more likely to trust dashboards when they can see exactly how numbers are calculated. Make the data dictionary available to everyone, and invite questions. If someone believes a number is wrong, there should be a clear way to check.
Signs you need to standardize
- Leaders spend meetings debating whose numbers are right.
- Two systems report different values for the same metric.
- A building looks unusually good or bad compared to peers without an operational explanation.
- Trends change abruptly when someone new takes over reporting.
Start with three
Do not attempt to standardize everything at once. Choose the three metrics you discuss most often and get them right. Expand from there.
CarePulse calculates group metrics from one set of definitions across all of your buildings, and shows those definitions alongside the numbers. If you would like to see how that works, we are glad to walk you through a demo.