Understanding and Using SPC Charts in Healthcare Operations

Why this topic matters

Operational teams in the NHS are constantly striving to improve efficiency, patient safety, and overall service delivery. However, making informed decisions about where and how to intervene can be challenging without a clear understanding of process performance. Traditional metrics often present data as averages or simple percentages, which can mask important underlying trends and variations.

SPC charts provide a robust, evidence-informed method to:

• Distinguish signal from noise: Identify when a process is truly changing versus when it is merely exhibiting expected random variation.
• Focus improvement efforts: Direct resources towards addressing special causes when they occur or redesigning the system when common cause variation is too wide.
• Monitor the impact of change: Assess whether an improvement intervention has led to a statistically significant and sustained improvement.
• Promote data-driven decision making: Move beyond anecdotal evidence to make decisions based on robust statistical analysis.

Practical explanation: Common and Special Cause Variation

At the heart of SPC charts is the concept of variation. Every process, whether clinical or administrative, exhibits variation. Understanding the nature of this variation is crucial for effective improvement.

Common Cause Variation

Common cause variation, also known as 'random' or 'unassignable' variation, is inherent to the process itself. It's the small, unpredictable fluctuations that occur even when a process is operating stably and predictably. Think of the slight differences in blood pressure readings from the same patient taken minutes apart, or the minor variations in patient waiting times for a clinic appointment on different days, assuming no major external disruptions.

When a process exhibits only common cause variation, it is said to be 'in statistical control' or 'stable'. While this doesn't necessarily mean the process is performing optimally, it does mean its performance is predictable within certain limits. To improve a process with only common cause variation, the system itself must be redesigned or fundamentally altered.

Special Cause Variation

Special cause variation, also known as 'assignable' variation, arises from specific, identifiable events or circumstances that are not part of the process's usual operation. These are often external factors or sudden changes. Examples include a new piece of equipment malfunctioning, a staff shortage due to an unexpected illness, or a sudden surge in patient demand due to a public health event.

When special cause variation is present, the process is considered 'out of statistical control' or 'unstable'. It is unpredictable, and its performance cannot be relied upon. The focus for improvement here is to investigate and eliminate (or, if beneficial, embed) the specific cause.

The Role of Control Limits

SPC charts plot data points over time, and critically, they include 'control limits'. These are statistically derived lines that define the expected range of common cause variation. They are typically set at three standard deviations above and below the process average.

• Upper Control Limit (UCL): The upper boundary for expected common cause variation.
• Lower Control Limit (LCL): The lower boundary for expected common cause variation.
• Centre Line (CL): Represents the average performance of the process.

When a data point falls outside these control limits, or when specific patterns are observed within the limits (e.g., several points in a row trending upwards), it signals the presence of special cause variation.

Step-by-step approach to using SPC Charts

1. Define the Process and Measure

Clearly identify the process you want to monitor and define the measure you will track. This should be a meaningful operational metric, e.g., 'time from decision to admit to ward bed', 'percentage of clinical notes completed within 24 hours', or 'number of patient falls per 1000 bed days'. Ensure the data can be collected consistently and reliably.

2. Collect Data

Gather sufficient data points over time. For initial chart construction, typically 20-25 data points are recommended to establish a baseline. Continue collecting data regularly as the process unfolds.

3. Choose the Right SPC Chart Type

There are various types of SPC charts, each suited to different types of data:

• XmR (Individual and Moving Range) Chart: For continuous data where individual data points are collected (e.g., waiting times, patient age).
• np Chart: For count data where the sample size is constant (e.g., number of patients with healthcare-associated infections in fixed batches of 100).
• p Chart: For proportion or percentage data where the sample size varies (e.g., percentage of appointments attended out of total scheduled, where the total scheduled varies daily).
• c Chart: For count data of defects per unit where the 'unit' is constant (e.g., number of prescribing errors per 100 prescriptions).
• u Chart: For count data of defects per unit where the 'unit' varies (e.g., number of pressure ulcers per 1000 inpatient days).

Selecting the correct chart is critical for accurate analysis. Resources from NHS England or the Institute for Healthcare Improvement (IHI) can provide further guidance on chart selection.

4. Calculate Control Limits

Based on your collected baseline data, calculate the centre line, UCL, and LCL. Most QI software or even advanced spreadsheet functions can perform these calculations automatically. The calculations are based on the process average and standard deviation (or proportion/count equivalents).

5. Plot the Data and Interpret the Chart

Plot your data points sequentially over time, along with the centre line and control limits. Look for 'rules' that indicate special cause variation. Common rules include:

• Points outside control limits: Any single point above the UCL or below the LCL.
• Runs: Seven or more consecutive points all above or all below the centre line.
• Trends: Seven or more consecutive points all steadily increasing or decreasing.
• Patterns: Unusual patterns or non-random distribution of points.

6. Act on the Signals

• If special cause variation is detected: Investigate the specific event or change that led to the signal. If it's an undesirable special cause (e.g., an increase in adverse events), take action to eliminate it. If it's a desirable special cause (e.g., a sustained drop in waiting times after an intervention), understand what caused it and embed the change.
• If only common cause variation is present: The process is stable but its average performance may not be good enough. To improve, you need to change the fundamental design of the process, not just react to individual points.

Common pitfalls

• Reacting to common cause variation: Mistaking random fluctuations for meaningful changes, leading to unnecessary interventions that destabilise the process.
• Incorrect chart selection: Using the wrong type of SPC chart for the data can lead to erroneous conclusions about process control.
• Insufficient data: Not having enough baseline data to establish reliable control limits, or stopping data collection too early.
• Ignoring the context: Interpreting charts purely statistically without understanding the operational context and what might have influenced the data points.
• Lack of team engagement: SPC charts are most effective when the team involved in the process understands and uses them, fostering a culture of continuous improvement.

Example in clinical practice: Monitoring discharge summary completion

A hospital trust wants to improve the timely completion of discharge summaries for patients, aiming for 95% to be completed within 24 hours of discharge. The operational team decides to monitor the 'percentage of discharge summaries completed within 24 hours' using a p-chart, as the total number of discharges (sample size) varies daily.

1. Data Collection: They collect daily data for 30 days, recording the number of discharges and the number of summaries completed within 24 hours.
2. Chart Construction: Using this baseline data, they calculate the overall average percentage (centre line) and the upper and lower control limits.
3. Monitoring: Each day, they plot the new percentage on the chart.
4. Interpretation:
   • For the first few weeks, the points fluctuate within the control limits. This indicates the process is 'in control' but its average performance is only 85%, consistently below the 95% target. This signals that fundamental system changes are needed.
   • After implementing a new electronic discharge system and staff training, a point falls above the UCL, followed by a run of seven points all above the previous centre line. This is a special cause signal, indicating a statistically significant improvement. The team investigates to confirm the new system is the cause and decides to recalculate control limits based on the improved performance.
   • Later, a point falls below the LCL. The team investigates and discovers a significant IT system outage on that day prevented staff from accessing the summary system, a clear special cause to address with IT.

This example illustrates how SPC charts help differentiate between expected variation and significant changes, guiding targeted actions rather than reactive firefighting.

How Lazomis can help

Lazomis provides integrated tools that simplify the creation and interpretation of SPC charts. Our platform allows you to:

• Effortlessly import and visualise data: Connect your operational data sources or manually input data to automatically generate a variety of SPC charts, including XmR, p, np, c, and u charts.
• Automated control limit calculation: Reduce the burden of manual calculations, ensuring accuracy in your charts.
• Highlight special cause signals: Our charts automatically flag common rules for special cause variation, helping your team quickly identify when interventions are needed.
• Support collaborative improvement: Share charts easily with your team, annotate specific events, and track improvement efforts within the Lazomis environment.
• Integrate with QI projects: Embed SPC charts directly into your Lazomis QI projects, providing clear visual evidence of progress and impact.

This resource supports, but does not replace, clinical judgement. Local policy, formulary and specialist advice should be followed.

Key takeaways