Module 6 — How to Read a Dashboard

What is the dashboard really telling us — and what might we misunderstand?

Module Learning Objective

This module helps explain:

how healthcare dashboards should be interpreted critically — and why dashboards are often starting points for questions rather than endpoints for conclusions.

By the end of this module, readers should feel more confident asking:

What is this dashboard really telling us — and what might I be misunderstanding?

The module focuses on practical healthcare dashboards commonly used to monitor:

Emergency Department (ED) activity
emergency admissions
elective recovery
ambulance handover delays
delayed discharge
pathway performance
population health and inequalities
operational flow

Rather than teaching how dashboards are built, this module focuses on:

how dashboards should be interpreted

to support better questions and better decision-making.

Why Dashboards Matter

Healthcare systems increasingly rely on dashboards.

Every day, leaders across healthcare review information intended to support decisions about:

demand
operational performance
quality
access
inequalities
workforce pressure
system improvement

Examples may include:

Emergency Department performance dashboards
elective waiting list dashboards
ambulance handover dashboards
discharge and flow dashboards
population health dashboards
inequalities dashboards
pathway monitoring dashboards

Dashboards are powerful because they help simplify large volumes of information.

Instead of reading lengthy reports:

decision-makers can quickly see:

performance trends
variation
comparison
risk
areas requiring attention

However:

dashboards do not speak for themselves

A dashboard presents information.

It does not automatically explain:

why something happened

what should happen next

This distinction matters.

Because poor dashboard interpretation can lead to:

overreaction to normal variation
misleading conclusions
poor investment decisions
misplaced operational attention
false reassurance

Good dashboard interpretation therefore asks:

What am I really looking at?

before asking:

What should we do?

A Practical Healthcare Example

Imagine an executive team reviewing a monthly urgent and emergency care dashboard.

The dashboard shows:

Measure	Last Month	This Month
ED attendances	18,400	20,100
4-hour performance	76%	69%
Ambulance handover delays	420	590
Emergency admissions	5,230	5,260

At first glance:

this looks concerning.

A quick conclusion might be:

Urgent care performance is deteriorating.

But an important question remains:

What is actually driving this change?

Before acting, a decision-maker might reasonably ask:

Is this seasonal?
Is this normal variation?
Are attendances increasing because of respiratory illness?
Is performance deteriorating everywhere nationally?
Is there a meaningful trend — or a single-month fluctuation?
Are the metrics connected?

This illustrates an important idea:

dashboards rarely provide answers on their own

Instead:

they help us ask better questions

Common Dashboard Interpretation Mistakes

One of the biggest risks in healthcare analytics is assuming dashboards are self-explanatory.

In practice:

well-intentioned decision-makers often draw conclusions too quickly.

The goal is not to criticise dashboards.

Instead:

to avoid drawing the wrong conclusion too quickly

Mistake 1 — Focusing Only on Red, Amber and Green Status

Many dashboards use:

🔴 Red
🟠 Amber
🟢 Green

to summarise performance.

This can be useful.

But it can also oversimplify reality.

For example:

A metric may appear:

🔴 RED

because a target was narrowly missed.

Meanwhile:

performance may actually be improving.

Similarly:

something shown as:

🟢 GREEN

may still hide important problems.

For example:

performance may technically exceed threshold —

but still be deteriorating over time.

A useful question becomes:

Compared with what?

For example:

previous performance?
peer organisations?
national trend?
expected seasonal variation?

Good dashboard interpretation asks:

What sits behind the colour?

rather than simply:

Is it red or green?

Mistake 2 — Confusing Counts with Rates

This links directly back to:

Module 1 — Counts vs Rates

A dashboard may show:

Higher emergency admissions

But this does not automatically mean:

worse performance.

Why?

Because larger populations naturally generate larger numbers.

For example:

an Integrated Care Board serving:

1 million people

would reasonably be expected to have:

more admissions

than one serving:

350,000 people

Similarly:

areas with older populations may naturally experience:

higher healthcare utilisation.

This means dashboard interpretation should ask:

Am I looking at counts or rates?

and:

Is this comparison fair?

For example:

crude counts?
rates per population?
age-standardised comparison?
deprivation-adjusted comparison?

Without context:

dashboards can unintentionally mislead.

Mistake 3 — Overreacting to Single-Month Changes

Healthcare systems naturally fluctuate.

This links directly back to:

Module 4 — Variation & Distributions

For example:

imagine:

ED attendances ↑ 8%

in one month.

Question:

Does this represent a meaningful change?

Maybe.

But maybe not.

Healthcare systems are noisy.

Variation happens because of:

seasonal illness
bank holidays
staff shortages
coding differences
operational disruption
random fluctuation

A single month of deterioration does not necessarily mean:

performance is collapsing.

Similarly:

a single month of improvement does not necessarily mean:

a scheme worked.

This is why dashboards often benefit from:

trends

rather than isolated points.

Or:

Statistical Process Control (SPC)

which helps distinguish:

normal variation

from

meaningful change

A useful question becomes:

Are we observing signal — or noise?

The image below illustrates an important distinction between:

reacting to dashboards

and

interpreting dashboards critically

Fast dashboard interpretation can sometimes encourage conclusions based on colours, isolated metrics or single-month changes.

More thoughtful interpretation instead asks:

What context, comparison or explanation might sit behind the numbers?

Good dashboard interpretation focuses less on immediate conclusions and more on:

context, comparison, variation and critical questioning

Dashboards help surface patterns and areas requiring attention.

However:

interpretation still requires context, judgement and questioning

Before moving on, consider:

Am I reacting to the dashboard — or interpreting it?

Mistake 4 — Ignoring Denominators

Another common mistake is focusing only on raw numbers.

A dashboard may show:

50 emergency admissions from care homes

Question:

Is this high or low?

The answer depends on:

how many care home residents exist
the size of the older population
expected baseline utilisation
comparison with peers

Without this context:

raw numbers become difficult to interpret.

Similarly:

40 inpatient falls

might initially feel alarming.

But interpretation changes dramatically if this occurred across:

500 admissions

versus:

25,000 occupied bed days

This matters for:

admissions
complaints
mortality
harm incidents
screening uptake
readmissions
pathway conversion

Raw numbers alone rarely tell the full story.

Good dashboard interpretation therefore asks:

Relative to what?

Mistake 5 — Assuming Correlation Means Causation

This links back to:

Module 3 — Correlation vs Causation

Dashboards frequently show things moving together.

For example:

ED attendance ↑
4-hour performance ↓

Question:

Did one cause the other?

Maybe.

But dashboards alone rarely prove causation.

Other factors may contribute:

staffing
flow delays
bed occupancy
discharge delays
seasonal illness

A dashboard may show:

association

without proving:

cause

Good interpretation asks:

What else might explain this pattern?

Reflection

Before moving on, ask:

If a dashboard suddenly looked concerning, what questions would I want answered before acting?

Good dashboard interpretation begins with curiosity rather than immediate conclusions.

Dashboards become most useful when they encourage:

better questions rather than faster conclusions

rather than immediate conclusions.

Good dashboard interpretation is not about memorising metrics.

Instead:

it is about learning how to think critically about what the dashboard might — and might not — be telling us

The following questions can provide a practical framework for interpreting dashboards more effectively.

1. What Am I Actually Looking At?

Before interpreting a metric, it helps to understand:

what the measure actually represents

For example:

is it:

a count?
a rate?
a percentage?
an average?
a median?
a target?
a trend?

This matters.

A dashboard may show:

Emergency admissions ↑

Question:

Compared with what definition?

Does this include:

same-day emergency care?
observation activity?
short stay admissions?
only non-elective inpatient spells?

Similarly:

a dashboard may show:

Average waiting time

Question:

Average according to whom?

Mean?

Median?

90th percentile?

The same label can sometimes describe very different things.

A useful first question therefore becomes:

What exactly am I looking at?

2. Compared With What?

Numbers rarely mean much in isolation.

A dashboard showing:

72%

immediately prompts another question:

Compared with what?

Possible comparisons include:

previous month
previous year
peer organisations
national average
target or threshold
expected seasonal pattern
statistical control limits

For example:

ED performance may appear poor.

However:

if all peer systems experienced winter deterioration during the same period:

the interpretation changes.

Similarly:

an increase in emergency admissions may feel alarming.

But interpretation changes if:

demand increased nationally

or:

population need increased substantially.

Good dashboard interpretation asks:

Is this number unusual — or expected?

3. Is This Meaningful Change — or Normal Variation?

Healthcare systems fluctuate.

Not every change matters.

A common mistake is assuming:

performance changed

therefore:

something important happened

Sometimes this is true.

Sometimes it is simply:

normal variation

For example:

a dashboard may show:

ED attendances ↑ 4%

Question:

Is this meaningful?

Good interpretation asks:

Is this one month or a sustained trend?
Does the change exceed normal fluctuation?
What does the SPC chart suggest?
Has seasonality been considered?

The key question becomes:

Are we observing signal — or noise?

The framework below summarises a practical way to interpret dashboards before jumping to conclusions.

Rather than reacting immediately to a metric or colour:

pause, question and interpret.

4. What Context Might Explain This?

Dashboards rarely explain:

why

something happened.

Context matters.

For example:

a deterioration in urgent care performance may reflect:

winter respiratory illness
workforce shortages
discharge delays
bed occupancy
industrial action
increased acuity
coding or operational changes

Similarly:

an apparent improvement may not always mean:

improvement caused by intervention.

Good interpretation therefore asks:

What else changed at the same time?

This links closely with:

Module 7 — Thinking About Evidence & Evaluation

because dashboard interpretation should avoid:

jumping from association to causation.

5. What Decision Does This Information Support?

Perhaps the most important question is:

What decision are we trying to make?

Not every metric needs action.

Not every deterioration requires escalation.

And not every improvement should trigger celebration.

A dashboard should help support:

prioritisation
investigation
monitoring
resource allocation
quality improvement
escalation where appropriate

Good interpretation therefore asks:

What decision does this information meaningfully support?

Sometimes the right answer is:

monitor more closely

rather than:

act immediately

Reflection

Before acting on dashboard information, pause and ask:

What am I actually looking at, compared with what, and what context might explain it?

Good dashboard interpretation moves from:

seeing numbers

thinking critically about meaning

A Simple Dashboard Interpretation Checklist

Before acting, ask:

What am I looking at?
Compared with what?
Signal or noise?
What context matters?
What decision does this support?

Key Takeaways

dashboards present information — they do not automatically explain it
numbers rarely make sense without context or comparison
variation matters — not every change is meaningful
dashboards show patterns but rarely prove causation
good interpretation focuses on questioning before action

Effective dashboard interpretation moves beyond:

What happened?

to ask:

What might explain this — and what should we do next?

Questions Decision-Makers Should Ask

Before acting on dashboard information, ask:

What am I actually looking at?
Compared with what?
Is this signal or normal variation?
What context might explain this?
What decision does this information support?
What additional information would improve confidence?

Dashboards are most valuable not when they provide answers — but when they help decision-makers ask better questions.