Section 7.2: Statistical Process Control & Control Charts

Statistical Process Control & Control Charts

Statistical Process Control (SPC) is a branch of quality science that uses statistical methods to monitor, control, and improve processes. Developed by Walter Shewhart at Bell Laboratories in the 1920s and popularized by W. Edwards Deming, SPC provides healthcare quality professionals with a rigorous, mathematical framework to distinguish between expected, everyday variation and unusual, system-level disruptions. By utilizing control charts, organizations can avoid the costly mistake of reacting to normal fluctuations or failing to intervene when a process is truly unstable.

Common Cause vs. Special Cause Variation

At the heart of SPC is the distinction between two types of process variation. Understanding this difference is one of the most frequently tested concepts on the CPHQ exam.

1. Common Cause Variation

Common cause variation (also called random or systemic variation) is inherent in the design of the process. It is the cumulative effect of many small, unavoidable causes.

• Characteristics: It is stable, predictable within limits, and consistent over time.
• Healthcare Example: The slight variation in laboratory test turnaround times (e.g., ranging between 35 and 48 minutes) due to minor differences in walking speed, specimen loading times, or normal machine processing cycles.
• Management Strategy: Because common cause variation is built into the system, frontline staff cannot eliminate it through local adjustments. To reduce common cause variation, leadership must redesign the system itself (e.g., purchasing automated equipment or revising the entire workflow). Reacting to individual common-cause data points as if they are unique events is called tampering or overadjustment, which actually increases overall process variation and wastes resources.

2. Special Cause Variation

Special cause variation (also called non-random or assignable variation) arises from factors outside the normal design of the process.

• Characteristics: It is unstable, unpredictable, and represents a shift in process performance.
• Healthcare Example: A sudden spike in laboratory turnaround times to 120 minutes because a centrifuge broke down or a new, untrained technician was assigned to the shift.
• Management Strategy: Special cause variation requires immediate investigation. Frontline teams must identify the root cause (the 'assignable cause') and implement local, corrective action to return the process to its normal state.

The Dangers of Overadjustment (Tampering)

Deming's Funnel Experiment illustrates the mathematical consequences of tampering. In this experiment, a marble is dropped through a funnel aimed at a target. If the operator adjusts the funnel position after each drop to compensate for the distance from the target, the variance of the drops actually doubles compared to leaving the funnel in a fixed position. In healthcare, this translates to clinicians changing protocols after a single bad outcome, which introduces more chaos and instability into the system rather than improving it.

---

Anatomy of a Control Chart

A control chart is a run chart with mathematically calculated limits. It consists of:

• Center Line (CL): Represents the process average (the mean, $\bar{X}$).
• Upper Control Limit (UCL): Positioned exactly three standard deviations ($+3\sigma$) above the mean.
• Lower Control Limit (LCL): Positioned exactly three standard deviations ($-3\sigma$) below the mean.

Control limits represent the boundaries of common cause variation. As long as all data points remain within these limits and fluctuate randomly around the mean, the process is said to be in statistical control.

[!IMPORTANT] Control limits are calculated using the process's historical performance data. They are entirely separate from specification limits or clinical targets, which are external goals set by management, regulators, or clinical guidelines. A process can be in statistical control (stable and predictable) but fail to meet clinical specifications (e.g., a stable infection rate that is higher than the national target).

---

The Western Electric Rules

To identify special cause variation, quality professionals divide the space between the mean and the control limits into three zones, each one standard deviation ($1\sigma$) wide:

• Zone C: Within $1\sigma$ of the center line.
• Zone B: Between $1\sigma$ and $2\sigma$ from the center line.
• Zone A: Between $2\sigma$ and $3\sigma$ from the center line.

A process is unstable (out of control) if any of the following Western Electric rules are triggered on either side of the mean:

1. Rule 1 (Outlier): A single data point falls outside the 3-sigma limits (above the UCL or below the LCL).
2. Rule 2 (Zone A): Two out of three consecutive points fall in Zone A or beyond (more than $2\sigma$ from the mean, on the same side).
3. Rule 3 (Zone B): Four out of five consecutive points fall in Zone B or beyond (more than $1\sigma$ from the mean, on the same side).
4. Rule 4 (Shift): Eight consecutive data points fall on one side of the center line (in Zone C or beyond).

---

Control Chart Selection Matrix

Selecting the correct control chart is a critical CPHQ skill. The choice depends on the type of data (attribute vs. variable) and the subgroup characteristics.

Attribute vs. Variable Data Rules

• Defective vs. Defects: A 'defective' refers to an entire unit that is unacceptable (e.g., an infected patient). A 'defect' refers to an individual flaw on a unit (e.g., a patient could have three separate pressure ulcers; that represents three defects on one defective patient).
• X-bar and R Charts: These are used together. The X-bar chart monitors the process average (mean), while the R (Range) chart monitors process variability (range of the subgroup). If process variability increases, the control limits on the X-bar chart will widen, making the chart less sensitive to shifts. Therefore, variability must be stabilized first.