Phase 1: Collect and Visualize Your Data
The first phase is to collect and visualize your data to establish a baseline for analysis. The initial tool you will create is a simple spreadsheet with two columns for the date and the fall rate value over at least 12-20 months. You will then input this data into an analysis platform to generate the primary visual tool for this workflow: a run chart. This chart plots your data points over time and includes a median line, making it easy to spot potential patterns visually.
What would you like me to fetch?
1. Collect and Visualize Your Data
Your first step is to gather the historical data and plot it over time to see what it looks like. This initial visualization is the foundation for all further analysis.
Action Plan
- Gather Your Data: Collect at least 12-20 data points of your hospital’s fall rate. The most common metric is falls per 1,000 patient days. Your Health Information Management or Decision Support department can likely provide this data.
- Format Your Data: Create a simple table or spreadsheet. You only need two columns: the Date (e.g., Jan 2024, Feb 2024) and the Value (the fall rate for that month).
Month | Falls per 1,000 Patient Days |
Jan 2024 | 2.5 |
Feb 2024 | 2.8 |
Mar 2024 | 2.3 |
- Create a Run Chart: Input this data into your analysis platform to generate a run chart. A run chart is simply a line graph that shows your data points over time with a center line, which is the median. This chart makes it easy to spot potential patterns visually.
1. Collect and Visualize Your Data
Collecting Data and Creating a Run Chart
Phase 2: Identify Trends or Shifts
Next, the analysis phase uses statistical rules to determine if the patterns on your run chart are meaningful. The core tool here is your analysis platform’s automated application of run chart rules, which looks for non-random signals like a “shift” (six points below the median) or a “trend” (five points all going up or down). If no signal is found, the run chart itself becomes a strategic discussion tool to help leadership decide if redesigning a stable process is a priority. The platform may also serve as a tool by recommending a more appropriate chart type, such as a T-Chart, for tracking rare events.
What would you like me to fetch?
2. Apply Rules to Identify Trends or Shifts
Now, you’ll use specific statistical rules to determine if the patterns you see on the chart are just random noise or a meaningful signal that something has changed. This is the key to avoiding guesswork. Your platform should automate this analysis.
Action Plan
- Analyze the Run Chart: The platform will automatically apply the run chart rules to look for non-random patterns indicating a real change in performance. The four main signals are:
- Shift: Six or more consecutive points are all above or all below the median. This suggests a sustained change has occurred.
- Trend: Five or more consecutive points are all going up or all going down. This indicates a gradual change in performance.
- Too Few or Too Many Runs: A “run” is a series of one or more consecutive points on the same side of the median. If there are too few or too many runs, it suggests the variation is not random.
- Astronomical Point: A single data point that is clearly an outlier, far from the others.
- Select the Right Chart: A good platform will also suggest if a different type of chart is more appropriate. For example, if falls are a rare event at your hospital, it might recommend a T-Chart, which measures the “days between” falls. This can be more sensitive for tracking rare events.
2. Identify Trends or Shifts
A Guide to Interpreting Run Charts_ The Four Rules for Finding Non-Random Signals
Phase 3: Focus the QI Team's Efforts
Finally, if a signal is detected, the third phase focuses on drilling down to generate actionable insights for a quality improvement team. Here, you will use the analysis platform’s data stratification tools to re-run the analysis and create separate run charts for different subgroups, such as by hospital unit, time of day, or type of fall. The ultimate output is an automated report generated by the platform, which can provide a specific insight like, “A negative trend in fall rates began in July 2024 and is isolated to the 3 East unit during the night shift”. This report becomes the tool that gives the QI team a focused, data-backed charter instead of a vague goal.
What would you like me to fetch?
Interpret the Results: Is a New Team Needed?
The results from your run chart analysis give you a clear, data-driven starting point for a conversation with leadership.
Scenario A: No Signal is Detected
If the platform’s analysis finds no trends or shifts, your process is considered stable and in statistical control. This doesn’t mean the fall rate is good, only that it’s predictable.
- Workflow Trigger: The absence of a signal is your cue to meet with the Chief Nursing Officer (CNO).
- Discussion Points: Present the run chart showing a stable process. Frame the conversation strategically:
- “The data shows our fall rate has been stable and predictable for the past 20 months. While we always want zero falls, the process isn’t actively getting worse.”
- “This stability suggests that a new QI team would be focused on redesigning a fundamentally stable process, which is a significant undertaking.”
- “Given our limited resources, should we confirm that redesigning this stable process is a higher priority than addressing other areas that may have negative trends?”
This approach uses data to facilitate a high-level strategic discussion about resource allocation.
Dig Deeper: Focus the QI Team's Efforts
Scenario B: A Signal is Detected
If the platform identifies a trend or a shift, you have found special cause variation. This is a clear signal that something specific has changed, and it provides an immediate focus for your new QI team.
- Drill-Down Analysis: Use the platform’s tools to stratify your data. Re-run the analysis by breaking down the data to find where the signal is coming from. Ask the platform to generate separate run charts for:
- Hospital Unit: Is the trend only happening on the medical-surgical unit?
- Time of Day: Are more falls happening during the night shift?
- Day of the Week: Is there a spike in falls on weekends?
- Type of Fall: Are the falls primarily from the bed, or are they ambulatory?
- Generate Actionable Insights: The platform should help you interpret these drilled-down charts. For example, it might generate a report stating: “A negative trend (shift) in fall rates began in July 2024 and is isolated to the 3 East unit during the night shift.”
- Inform the Team: This specific, data-backed insight is the perfect starting point for your QI team. Instead of a vague goal to “reduce falls,” their charter can be to “investigate and address the factors contributing to the increased fall rate on 3 East during night shifts since July.” This makes their work focused, measurable, and far more likely to succeed.
3. Focus the QI Team's Efforts
Worksheet A Signal is Detected!