EPISODE 4: "PDSA Cycles & Reading SPC Charts Without the Tears"

EPISODE 4: PDSA Cycles & Reading SPC Charts Without the Tears

Subtitle: The "Rule of 1-1-1," Why Control Limits Are Not Targets, and the 4 Rules of Special Cause Variation

Episode Overview

Why do traditional before-and-after bar charts and $p$-values lie in Quality Improvement? Because every Emergency Department has natural day-to-day up-and-down bounce. In Episode 4, we demystify rapid-cycle Plan-Do-Study-Act (PDSA) testing and teach you how to read Run Charts and Statistical Process Control (SPC / Shewhart) Charts in 30 seconds flat—including the exact 4 mathematical rules tested on the MRCEM exam.

High-Yield Clinical & Exam Takeaways

• The PDSA Cycle & The "Rule of 1-1-1": Why you must test your first prototype on 1 patient, with 1 nurse, on 1 afternoon shift—and how to decide whether to Adopt, Adapt, or Abandon after each cycle.
• Run Charts vs. SPC Charts: Understanding the Median vs. the Mean ($\mu$), and how Upper ($\text{UCL} = \mu + 3\sigma$) and Lower ($\text{LCL} = \mu - 3\sigma$) Control Limits are calculated from the inherent variability of the system itself.
• The #1 Exam Trap — Control Limits $\neq$ Clinical Targets: Why a process with a Mean of $110\text{ mins}$ and an $\text{LCL}$ of $75\text{ mins}$ is statistically stable ("in control")—yet 100% incapable of ever hitting a $60\text{ min}$ clinical target without fundamental system redesign.
• Common Cause Variation (Noise) vs. Special Cause Variation (Signal): Why reacting to single bad days inside the Control Limits ("Tampering") makes a stable system worse.
• The 4 Universal Rules for Detecting Special Cause Variation ($p < 0.05$ equivalent):

1. Rule 1 (Astronomical Point): 1 single point outside the $3\sigma$ Control Limits ($\text{UCL}$ or $\text{LCL}$).
2. Rule 2 (The Shift): $6\text{ to }8+$ consecutive points all strictly above or all strictly below the Center Line (even if they stay inside the Control Limits!).
3. Rule 3 (The Trend): $5\text{ to }6+$ consecutive points all steadily increasing or all steadily decreasing.
4. Rule 4 (Runs / Sawtooth Oscillation): Non-random zig-zagging across the Center Line (usually caused by mixing Day Shift and Night Shift data on the same chart!).

Clinical Vignettes Covered

• Safe Adult Procedural Sedation (3-Cycle PDSA Story): Watching compliance evolve from 37.5% (failed Cycle 1 paper forms in pigeonholes) $\rightarrow$ 71% (Cycle 2 EHR macro + Omnicell screen tag) $\rightarrow$ 96.7% (Cycle 3 pre-packed "Safe Sedation Grab-Box" with nasal $\text{EtCO}_2$ cannula on top of the Resus airway cart).
• Stroke Door-to-Needle & Sepsis Door-to-Antibiotic SPC Charts: Identifying an IT-crash outlier (Rule 1), a 6-week adoption trend (Rule 3), and a locked-in 9-week shift (Rule 2).

The 3:00 AM Resus Bay Bottom Line: Never present a QIP with a two-column "Before vs. After" bar chart. Plot your data weekly on a Run or SPC Chart, annotate your PDSA cycles with arrows, and prove your intervention worked by achieving an 8-point Shift below the baseline median.