No products in the cart.
MODULE 12: KPI & QUALITY IMPROVEMENT ANALYTICS
Section 4: Operational Dashboards for Leadership
From Spreadsheets to Strategy: The art of visualizing data to tell a compelling story, drive decision-making, and provide leadership with at-a-glance operational awareness.
SECTION 12.4
Operational Dashboards for Leadership
Translating Complex Data into Clear, Actionable Insights.
12.4.1 The “Why”: Beyond the Report, Beyond the Spreadsheet
In the preceding sections, you have become a master of measurement. You know how to define a SMART KPI, extract the data from the EHR, and analyze its validity. You can produce a spreadsheet with pinpoint accuracy that shows the median STAT medication turnaround time for the last quarter. This is the foundational work of an analyst. However, if your final product is that spreadsheet—a dense grid of numbers emailed to a busy director—you have failed. You have delivered data, but you have not delivered insight. You have created a report, but you have not created understanding.
The human brain is not wired to find meaning in a table of numbers. We are visual creatures. We are pattern-seekers. We understand information far more quickly and intuitively when it is presented visually. A single, well-designed chart can convey a message more powerfully and memorably than a hundred rows of data. This is the “why” of the operational dashboard. A dashboard is not just a collection of charts; it is a visual argument. It is a narrative, told with data, designed to provide at-a-glance awareness of performance, highlight trends and outliers, and facilitate rapid, informed decision-making.
For hospital leadership, time is the most precious and limited resource. A Pharmacy Director or a Chief Medical Officer does not have thirty minutes to decipher your spreadsheet. They have thirty seconds, between meetings, to look at a screen and answer three critical questions: “Are we winning or losing?”, “Where are the problems?”, and “Are our improvement efforts working?” A successful dashboard answers these questions instantly. It transforms you from a data provider into a storyteller and a strategist. It elevates your work from the back-room server to the C-suite monitor, making your analysis an indispensable tool for steering the organization.
This section is your masterclass in the principles of data visualization and dashboard design. This is not about learning a specific software tool (like Tableau or Power BI), but about learning the timeless principles that apply to any tool. You will learn to think like a graphic designer, a journalist, and a psychologist all at once. You will learn how to choose the right chart for the right data, how to use color and layout to guide the viewer’s eye, and how to eliminate the “chart junk” that obscures the message. Your clinical background gives you a unique advantage: you already understand the underlying context of the data. Now, you will learn the art of presenting that data in a way that is clear, compelling, and impossible to ignore.
Retail Pharmacist Analogy: The Car Dashboard
Think about the dashboard in your car. It is a perfect, real-world example of an effective operational dashboard. Imagine trying to drive your car if, instead of a dashboard, you were handed a detailed engineering spreadsheet every five seconds with hundreds of data points: engine RPM, oil pressure in PSI, coolant temperature in degrees Celsius, individual wheel speed, etc. It would be impossible. You would be so busy trying to find the important numbers that you would crash.
The car’s dashboard solves this problem through brilliant design:
- It’s Contextual: It doesn’t just show you “55.” It shows you “55 MPH” on a speedometer with clear markings for the speed limit, giving you immediate context.
- It Prioritizes KPIs: Of the thousands of things that could be measured, the dashboard focuses on a vital few: speed, fuel level, engine temperature, and RPM. These are the KPIs for a safe and effective journey.
- It Uses the Right Visualizations: Speed is shown on a dial (a radial gauge) because the rate of change is important. Fuel is shown on a simple bar gauge because the absolute level is what matters. You don’t need to know the exact GPM (gallons per minute) your engine is consuming.
- It Uses Color for Pre-attentive Alerting: You don’t need to read the temperature gauge. When the needle enters the red zone, the color alone tells you there’s a problem that requires immediate attention. The “check engine” light is an even more direct alert. These visual cues are processed “pre-attentively” by your brain, before you even consciously read the label.
- It’s Designed for its Audience and Task: The driver’s dashboard is different from the one a mechanic uses when they plug in a diagnostic computer. The mechanic needs the detailed spreadsheet with the fault codes and sensor readings. The driver needs high-level, at-a-glance awareness to make immediate decisions.
When you build a dashboard for your Pharmacy Director, you are designing a car dashboard, not the mechanic’s diagnostic tool. Your director is the driver. They need to understand the speed, fuel level, and critical warnings of their department at a glance so they can navigate the complexities of the hospital. Your job is to filter out the engineering-level detail and present the vital signs needed to steer the ship.
12.4.2 The Psychology of Perception: How We See and Understand Data
Before we can design effective dashboards, we must first understand the basics of human visual perception. Our brains are hardwired to process certain visual information with incredible speed and efficiency. By leveraging these innate abilities, we can design charts and dashboards that are not just aesthetically pleasing, but cognitively easy to understand. The pioneers of data visualization, like Jacques Bertin and William Cleveland, conducted research that identified a hierarchy of “visual variables” that our brains are best at decoding to compare quantities.
The Visual Hierarchy: What the Brain Does Best
When comparing numerical values, our brains are not created equal. We are exceptionally good at judging some visual properties and surprisingly bad at others. This hierarchy is the most important foundational concept in data visualization.
Cleveland & McGill’s Hierarchy of Visual Perception
1.
Position on a common scale: Most Accurate. (e.g., Scatter plots, Bar charts with a common baseline)
2.
Length: Highly Accurate. (e.g., Bar charts)
3.
Angle / Slope: Moderately Accurate. (e.g., Line charts, Pie charts)
4.
Area: Poorly Accurate. (e.g., Bubble charts, Treemaps)
5.
Volume / Shading / Color Saturation: Least Accurate. (e.g., 3D charts, Heatmaps)
The Case Against Pie Charts
The hierarchy of perception explains why data visualization experts almost universally advise against using pie charts. A pie chart asks the viewer to compare quantities by judging angles and area (ranked #3 and #4 in accuracy). This is a task our brains are notoriously bad at. It is very difficult to tell if a slice representing 22% is definitively larger than a slice representing 20%.
A simple bar chart, which uses length on a common scale (#1 and #2), makes this comparison effortless and instantly accurate. As a general rule, if you are tempted to make a pie chart, make a bar chart instead. The only exception is when you are showing a part-to-whole relationship with only two or three distinct categories.
Pre-attentive Attributes: The Analyst’s Superpower
Pre-attentive attributes are the visual properties that our brains process in milliseconds, before we pay conscious attention to an object. By using these attributes strategically, you can guide your audience’s attention to the most important parts of your dashboard without them even realizing it.
Masterclass Table: Using Pre-attentive Attributes in Dashboard Design
| Attribute | Description | Strategic Application in a Pharmacy Dashboard |
|---|---|---|
| Color (Hue & Intensity) | A distinct color or a brighter/darker shade of the same color will “pop” out from its surroundings. | Use a neutral color (like gray or light blue) for the majority of your data. Reserve a single, bright, alerting color (like red or orange) to highlight only the data points that are failing to meet a target or represent a safety concern. This immediately draws the eye to the problem. |
| Size | An object that is noticeably larger than the others will command attention. | In a dashboard displaying KPIs, the most important number (e.g., “Preventable ADEs this month”) can be displayed in a much larger font size than the other metrics. This is often called a “Big Ass Number” (BAN) in design terminology. |
| Shape | A unique shape among a group of common shapes is instantly noticeable. | In a run chart showing daily performance, you could use a standard circle for each day, but use a distinct symbol (like a star or triangle) to mark the day a quality improvement intervention (a PDSA test) was implemented. |
| Position | We naturally read dashboards like we read a book: from top-left to bottom-right. The top-left quadrant is the most valuable real estate. | Place your highest-level, most important KPI (the one your Director cares about most) in the top-left corner of the dashboard. Place supporting details and more granular charts in the lower-right. |
12.4.3 The Dashboard Designer’s Playbook: Principles and Pitfalls
Armed with an understanding of human perception, we can now establish a set of core principles for effective dashboard design. Following these rules will help you create visualizations that are clear, concise, and impactful. Ignoring them is the fastest way to create a cluttered, confusing “data dump” that no one will use.
Core Principles of Effective Dashboard Design
The Five-Second Rule
A well-designed dashboard should allow a user to understand the general state of affairs—what’s good, what’s bad, and what’s changed—within five seconds of looking at it. If it takes longer than that to get the main message, the design has failed. This rule forces you to simplify, prioritize, and eliminate clutter.
Masterclass Table: Design Principles in Practice
| Principle | Explanation | Good Example vs. Bad Example |
|---|---|---|
| Choose the Right Chart for the Job | The type of data you have dictates the chart you should use. Time-series data needs a line chart. Categorical comparisons need a bar chart. Distributions need a histogram. | Good: A line chart showing the BCMA compliance rate over the last 12 months.
Bad: A pie chart showing the BCMA compliance rate for each of the last 12 months. |
| Maximize the Data-Ink Ratio | Coined by Edward Tufte, this principle states that a large share of the “ink” on a graphic should be dedicated to presenting the data itself. Erase non-data ink (like heavy gridlines, borders, unnecessary labels, 3D effects) whenever possible. | Good: A simple bar chart with light gray, thin gridlines or no gridlines at all.
Bad: A 3D bar chart with a dark background, thick white gridlines, and a border around the plot area. |
| Provide Context | A number without context is meaningless. Every key metric on a dashboard should be compared to something: a target, a historical average, a benchmark, a prior period. | Good: “STAT TAT: 22 min (Target: <20 min)”
Bad: “STAT TAT: 22 min” |
| Organize with a Visual Hierarchy | Use position, size, and color to guide the user. Place high-level summaries at the top and more detailed views at the bottom. The layout should flow logically. | Good: A dashboard with a row of KPI summary numbers at the top, followed by trend charts in the middle, and a detailed table at the bottom.
Bad: A jumble of charts of different sizes and types scattered randomly across the page. |
Dashboard Design Pitfalls: A Gallery of Horrors
The best way to learn good design is to see bad design. Below is a breakdown of common mistakes that can ruin an otherwise data-rich dashboard.
Example of a “Bad” Dashboard Design
ADC Stockouts by Drug (3D Pie)
Hard to compare slices. 3D effect distorts perception. What does this tell me to do?
BCMA Compliance (Tachometer)
A “tachometer” or gauge chart takes up huge space to show one number. Low data-ink ratio.
Turnaround Times for All Units
“Spaghetti graph” – too many lines (15 different units) on one chart makes it unreadable. Rainbow colors are distracting.
Analysis of Failures
- Wrong Chart Types: The 3D pie chart and gauge are poor choices for comparing data, as we’ve learned. They are classic examples of “chart junk.”
- Information Overload: The “spaghetti graph” tries to show everything at once and ends up communicating nothing. It would be far better to show a single line for the hospital-wide average and then use a separate bar chart to show the performance of each unit.
- Lack of Context: None of these charts show a target or a benchmark. Is 92% BCMA compliance good or bad? We have no idea.
- Poor Use of Color: The use of “rainbow” colors in the line chart is distracting and has no intrinsic meaning. Color should be used purposefully to encode information (e.g., gray for most units, red for the worst-performing unit).
12.4.4 A Blueprint for an Effective Pharmacy Leadership Dashboard
Now let’s apply these principles to build a conceptual model for a high-impact Pharmacy Leadership Dashboard. This blueprint is designed around the “Five-Second Rule” and the principle of organizing information in a logical hierarchy, from a high-level overview down to granular details. Most modern BI tools like Tableau or Power BI can be used to create a dashboard like this.
The Z-Layout: Guiding the Eye Naturally
The layout follows a natural “Z” pattern that mimics how we read a page. The eye starts at the top-left, scans across the top, moves diagonally down to the bottom-left, and then scans across the bottom. This places the most important information in the most prominent locations.
Conceptual Dashboard Layout
1. KPI Header: The “30,000 Foot View” (Top of the ‘Z’)
KPI 1
KPI 2
KPI 3
KPI 4
2. Main Trends: The “10,000 Foot View” (Middle of the ‘Z’)
Main Trend Chart 1
Main Trend Chart 2
3. Details & Drill-Down: The “1,000 Foot View” (Bottom of the ‘Z’)
Detail Chart 1
Detail Chart 2
Detail Table
Deconstructing the Layers
Let’s populate our blueprint with meaningful pharmacy KPIs, explaining the design choices for each layer.
Layer 1: The KPI Header Row
Purpose: To provide an immediate, at-a-glance summary of the most critical, high-level outcome metrics. This is what the C-suite cares about.
Design: Use “Big Ass Numbers” (BANs) with clear labels, context (targets), and trend indicators.
STAT TAT (Median)
22.5 min
+1.2%
Target: < 20 min
BCMA Compliance
98.2%
-0.5%
Target: > 98%
Drug Spend / APD
$215
-3.1%
Target: < $220
Preventable ADEs
2
(vs 2 prior mo)
Target: 0
Analysis of Design Choices
- Large Fonts (Size): The key number is the largest element, making it instantly readable.
- Strategic Color (Color): Performance below target (STAT TAT) is colored red. Performance at or above target is green. This provides instant good/bad judgment.
- Contextual Subtext: The target is clearly listed. The small percentage with an up/down arrow shows the trend vs. the prior period (e.g., last month), answering “Is it getting better or worse?”
Layer 2: The Main Trends
Purpose: To show the performance of key leading indicators over time. This layer helps leaders understand the story behind the KPI numbers and see if improvement efforts are having a sustained impact.
Design: Simple, clean line charts are the workhorse of this section.
Layer 3: The Details & Drill-Down
Purpose: To allow leaders to investigate the “why” behind a high-level trend. If the overall ADC stockout rate is high, this section answers the question, “Which cabinets and which drugs are the biggest problems?”
Design: Bar charts are excellent for comparing categories. Simple, clean tables are best for showing precise numerical data.
12.4.5 The Analyst as Storyteller: Building a Narrative with Data
The final and most advanced skill in dashboard design is moving from simply presenting data to telling a compelling story. A story has a structure, a narrative arc, and a clear point. Your dashboard should too. The goal is to guide your audience through the data in a logical way, leading them to the same conclusion you reached in your analysis.
The Pyramid Principle of Communication
Developed by consultant Barbara Minto, the Pyramid Principle is a powerful framework for executive communication. It states that you should start with the answer first. Give your main conclusion or key finding at the very beginning, and then present your supporting arguments and data in increasing levels of detail.
A great dashboard is a visual representation of the Pyramid Principle.
- The Point of the Pyramid (The Answer): The KPI Header Row. “Our STAT TAT is too high.”
- The Supporting Arguments: The Main Trend Charts. “It’s been trending up for three months.”
- The Foundational Data: The Detail Section. “The problem is primarily on the night shift and is worst on the surgical floors.”
The Power of Annotations and Titles
The fastest way to turn a chart into a story is to give it a title that states the main conclusion, and to add annotations that explain key events in the data.
A Bad Chart Title: “Monthly BCMA Compliance Rate”
A Good Storytelling Title: “BCMA Compliance Has Exceeded 98% Target Following Nurse Re-Education in August”
The second title doesn’t just describe the data; it interprets the data and presents a conclusion. It tells the reader what to look for and what it means.
From Static to Interactive: The Final Frontier
While a static, well-designed dashboard is a powerful tool, modern BI platforms allow for the creation of interactive dashboards that empower leaders to become their own analysts. Interactivity allows users to ask and answer their own questions by filtering, sorting, and drilling into the data.
Masterclass Table: Key Interactive Dashboard Features
| Feature | Description | Use Case for Pharmacy Director |
|---|---|---|
| Filters | Allow the user to slice the entire dashboard by a specific category, such as a nursing unit, a drug class, or a time of day. | The Director sees that the overall ADC stockout rate is high. They use a filter to select only the “ICU” units and instantly see that the ICU stockout rate is three times higher than the hospital average. |
| Tooltips | When a user hovers their mouse over a data point (like a bar on a bar chart), a small pop-up window appears with more detailed information. | The Director hovers over the bar representing the ICU stockouts and the tooltip shows the top 3 drugs causing those stockouts (e.g., Propofol, Fentanyl, Norepinephrine). |
| Drill-Downs (Hierarchies) | Allows the user to click on a high-level data point to “drill down” to the next level of detail. | The Director clicks on the “Opioids” bar in a “Drug Spend by Class” chart. The chart then re-draws to show the spend for each individual opioid (e.g., Hydromorphone, Fentanyl, Morphine), revealing that a new brand-name formulation is driving the cost increase. |
Building interactive dashboards is an advanced skill, but it is the ultimate goal for a mature informatics program. It transforms your role from a reactive report-runner into a proactive creator of analytical tools. By empowering your leadership with self-service analytics, you free up your own time to focus on the next level of complex problem-solving, creating a virtuous cycle of data-driven improvement for your organization.
