2.2 Quality Improvement, Patient Safety & Diagnostic Stewardship

Fundamentals of Healthcare Quality Improvement (QI)

Quality improvement (QI) in healthcare is the systematic, continuous application of structured methodologies to improve patient outcomes, system performance, and professional development. Clinicians must distinguish QI from traditional clinical research:

• Clinical Research: Aims to generate new, generalizable medical knowledge. It requires institutional review board (IRB) approval, rigid hypothesis testing, randomized control groups, and protocols that cannot be modified during the study.
• Quality Improvement: Aims to adapt existing evidence-based medical knowledge into local healthcare systems and clinical workflows. It relies on iterative, rapid-cycle testing, continuous protocol adaptation, and immediate local implementation.

The Institute of Medicine (IOM) Six Aims for Healthcare Improvement

In its landmark report Crossing the Quality Chasm, the IOM established six fundamental dimensions (the STEEEP criteria) that define high-quality healthcare:

1. Safe: Avoiding injuries to patients from the care that is intended to help them.
2. Timely: Reducing waits and harmful delays for both those who receive and those who give care.
3. Effective: Providing services based on scientific knowledge to all who could benefit, and refraining from providing services to those not likely to benefit (avoiding underuse and overuse).
4. Efficient: Avoiding waste, including waste of equipment, supplies, ideas, and human energy.
5. Equitable: Providing care that does not vary in quality because of personal characteristics such as gender, ethnicity, geographic location, or socioeconomic status.
6. Patient-Centered: Providing care that is respectful of and responsive to individual patient preferences, needs, and values, ensuring that patient values guide all clinical decisions.

---

The Model for Improvement & PDSA Cycles

Developed by the Associates in Process Improvement and widely adopted by the Institute for Healthcare Improvement (IHI), the Model for Improvement is the premier quality improvement framework in clinical medicine. It consists of two coordinated parts:

                  ┌──────────────────────────────────────────────┐
                  │            Part 1: Three Core Questions      │
                  │  1. What are we trying to accomplish? (Aim)  │
                  │  2. How do we know change is improvement?    │
                  │     (Measures: Outcome, Process, Balancing)  │
                  │  3. What change can we make? (Change Ideas)  │
                  └──────────────────────┬───────────────────────┘
                                         │
                                         ▼
                  ┌──────────────────────────────────────────────┐
                  │            Part 2: PDSA Testing Cycle        │
                  │          ┌───────► PLAN ────────┐            │
                  │          │                      │            │
                  │         ACT                    DO            │
                  │          │                      │            │
                  │          └─────── STUDY ◄───────┘            │
                  │        (Adapt, Adopt, or Abandon)            │
                  └──────────────────────────────────────────────┘

Part 1: The Three Foundational Questions

1. What are we trying to accomplish? (Setting the Aim): An effective Aim Statement must be SMART: Specific, Measurable, Achievable, Relevant, and Time-bound (e.g., "Increase the proportion of adult clinic patients with type 2 diabetes whose HbA1c is <8.0% from 54% to 75% within 12 months"). Broad, unquantifiable goals (e.g., "Improve diabetes care in our clinic") are ineffective.
2. How will we know that a change is an improvement? (Establishing Measures):
   • Outcome Measures: Direct reflection of the final clinical endpoint or patient health status (e.g., percentage of hypertensive patients with blood pressure <140/90 mm Hg, 30-day hospital readmission rates, colorectal cancer screening completion rates).
   • Process Measures: Quantify whether specific evidence-based clinical activities or workflow steps are being consistently performed (e.g., percentage of diabetic patients receiving annual microalbuminuria screening, percentage of patients with elevated initial blood pressure who receive a documented repeat blood pressure measurement after 5 minutes of rest).
   • Balancing Measures: Monitor for unintended negative consequences or trade-offs in other parts of the clinical system (e.g., tracking emergency department visits for severe hypoglycemia while driving intensive HbA1c reduction, or monitoring patient wait times when lengthening clinical consultation appointments).
   • Structural Measures: Reflect physical resources, staffing ratios, or EHR technology (e.g., ratio of registered nurse care managers to paneled patients, implementation of a clinical decision support alert).
3. What change can we make that will result in improvement? (Selecting Change Concepts): Formulating specific, evidence-based process modifications (e.g., introducing a standardized pre-visit planning huddle, creating standing orders for medical assistants to administer influenza vaccines).

Part 2: The Plan-Do-Study-Act (PDSA) Cycle

• Plan: Formulate the objective, develop explicit predictions of what will happen, and create a test plan (who, what, where, when, and how data will be gathered).
• Do: Carry out the test on a small scale (e.g., testing a new pre-visit checklist with 1 physician, 1 medical assistant, and 5 patients on a Tuesday morning). Document unexpected problems and observations.
• Study: Analyze the collected data, compare results against initial predictions, and synthesize key lessons learned.
• Act: Determine the next iterative step based on data:
  • Adapt: Modify the intervention to correct identified flaws and run a subsequent PDSA cycle.
  • Adopt: Implement the change on a larger scale across the entire clinic.
  • Abandon: Discard the change if data demonstrates lack of efficacy or severe adverse trade-offs, and develop a novel change concept.

Statistical Process Control: Run Charts vs. Control Charts

QI teams evaluate changes over time using dynamic graphical displays rather than simple static pre/post averages:

• Run Charts: Display data points plotted chronologically with a median line. Specific mathematical rules identify non-random "signals of change":
  • Shift: Six or more consecutive data points falling entirely above or below the median.
  • Trend: Five or more consecutive data points continually increasing or continually decreasing.
  • Astronomical Data Point: An extreme outlier that is blatantly obvious.
• Control Charts (Shewhart Charts): Plot data with a calculated center line (mean) and statistically derived Upper and Lower Control Limits (UCL and LCL, usually $\pm 3$ standard deviations). Distinguishes Common Cause Variation (inherent random system noise requiring fundamental system redesign) from Special Cause Variation (external, assignable perturbations requiring targeted investigation).

---

Lean, Six Sigma, and Ambulatory Process Optimization

Healthcare organizations frequently combine QI methodologies with industrial engineering frameworks:

Lean Methodology: Eliminating Waste

Originating from the Toyota Production System, Lean focuses on maximizing patient value by identifying and eliminating non-value-added waste (muda) from workflows. Value is defined strictly from the patient's perspective (activities that directly improve health or advance care):

• The 8 Healthcare Wastes (DOWNTIME Acronym):

  1. D — Defects: Medication errors, diagnostic mislabeling, missed laboratory results, incomplete clinical notes.
  2. O — Overproduction: Printing unused clinical forms, ordering screening tests more frequently than guideline intervals, redundant documentation.
  3. W — Waiting: Patients sitting in waiting rooms or exam rooms; physicians waiting for vitals, interpreters, or exam room turnover.
  4. N — Non-utilized Talent: Family physicians spending time on prior authorization phone calls or faxing records; nurses performing simple clerical tasks rather than triage and care management.
  5. T — Transportation: Physically moving medical charts, transport of lab specimens across distant facilities.
  6. I — Inventory: Storing excessive quantities of supplies, medications expiring on supply shelves before use.
  7. M — Motion: Clinicians and staff walking excessive distances across poorly designed clinic suites to retrieve tongue depressors, otoscope tips, or suture kits.
  8. E — Extra-Processing: Documenting the same clinical data in multiple disconnected EHR flowsheets, repetitive intake questionnaires.

• 5S Workplace Organization: A structured approach to organizing physical and digital workspaces:

  • Sort (Seiri): Remove unnecessary supplies and clutter from exam rooms.
  • Set in Order (Seiton): Arrange necessary tools in standardized, intuitive locations (e.g., standardizing the drawer layout of every exam room in the clinic).
  • Shine (Seiso): Clean, inspect, and maintain clinical equipment daily.
  • Standardize (Seiketsu): Establish uniform clinical protocols, visual cues, and labeling across all rooms.
  • Sustain (Shitsuke): Conduct regular audits to maintain discipline and habit.

Six Sigma Methodology: Reducing Process Variation

Originating from Motorola and General Electric, Six Sigma aims to reduce process variation and eliminate defects, targeting a defect rate of fewer than 3.4 defects per million opportunities ($99.99966%$ perfection). It utilizes the DMAIC framework:

• D — Define: Define the high-impact clinical problem and customer requirements.
• M — Measure: Measure baseline process performance and defect rates.
• A — Analyze: Analyze data to identify root causes of process defects and variation.
• I — Improve: Implement targeted solutions to eliminate root causes.
• C — Control: Establish standardized controls, dashboards, and audit procedures to sustain improvements.

---

Patient Safety Science & Root Cause Analysis (RCA)

The Swiss Cheese Model of System Accidents

Formulated by cognitive psychologist James Reason, the Swiss Cheese Model asserts that complex healthcare organizations possess multiple defensive layers designed to prevent errors (e.g., clinical protocols, barcode medication scanning, two-person verification, electronic decision support alerts). Each defensive layer, however, contains inherent vulnerabilities and weaknesses—analogous to the holes in slices of Swiss cheese.

• Latent Conditions: Organizational and systemic flaws that lie dormant for long periods (e.g., chronic understaffing, fatigued providers, confusing drug packaging, poorly configured EHR alerts, ambiguous handoff policies).
• Active Failures: Unsafe acts committed by individuals at the "sharp end" of care delivery (e.g., a nurse administering an incorrect vial, a physician overlooking an abnormal lab value).
• Adverse Event Alignment: An adverse event occurs only when the holes in multiple defensive slices momentarily align, allowing a latent hazard to pass through all defenses and cause patient harm. Focusing exclusively on punishing the individual at the sharp end fails to close the latent holes, ensuring the error will recur.

   Defensive Layers       (Swiss Cheese Slices)
   ┌────────┐    ┌────────┐    ┌────────┐    ┌────────┐
   │  Policy│    │Staffing│    │  EHR   │    │Physical│
   │   &    │    │   &    │    │ Alerts │    │ Look-  │
   │Protocol│    │ Fatigue│    │   &    │    │ Alike  │
   │        │    │        │    │ Usabil.│    │ Storage│
   └───┬────┘    └───┬────┘    └───┬────┘    └───┬────┘
       │             │             │             │
Hazard ◯─────────────◯─────────────◯─────────────◯──────► PATIENT
(Latent Holes Align Across All System Slices)              HARM!

Sentinel Events and Root Cause Analysis (RCA)

• Sentinel Event: Defined by The Joint Commission as a patient safety event (not primarily related to the natural course of the patient's illness) that reaches a patient and results in death, severe permanent harm, or severe temporary harm (e.g., wrong-site surgery, infant abduction, suicide in an inpatient facility, hemolytic transfusion reaction).
• Mandatory Response: Requires an immediate, comprehensive Root Cause Analysis (RCA).
• Core RCA Principles:
  • Retrospective Analysis: Examines an adverse event after it has occurred.
  • Non-Punitive & Systems-Focused: Focuses on identifying underlying vulnerabilities in processes, equipment, environment, and human factors, rather than assigning individual fault.
  • Multidisciplinary Team: Includes frontline clinicians, nurses, pharmacists, risk managers, and clinical technicians who understand the real-world workflow.

Key RCA Tools: The 5 Whys & The Ishikawa (Fishbone) Diagram

• The 5 Whys: Iteratively asking "Why?" (typically five times) to peel away proximate symptoms of a problem and expose the underlying organizational root cause.
  • Example: Why was an overdose of insulin administered? (Nurse selected wrong syringe). Why? (Standard U-100 syringes were out of stock). Why? (Supply cart was not restocked). Why? (Inventory clerk was reallocated to screening without coverage). Why? (No standardized cross-coverage staffing policy).
• Ishikawa (Fishbone / Cause-and-Effect) Diagram: Maps potential contributing factors into major categories:
  1. People: Staff fatigue, lack of training, communication breakdowns, supervision gaps.
  2. Process / Methods: Ambiguous protocols, lack of standardized handoffs, absent checklists.
  3. Equipment / Technology: EHR alert fatigue, look-alike pump interfaces, device malfunction.
  4. Materials: Look-alike/sound-alike medications, unstandardized supply bins.
  5. Environment: High noise levels, frequent clinical interruptions, poor lighting.
  6. Management / Policy: Chronic understaffing, production pressure, inadequate safety culture.

Proactive Risk Assessment: Failure Mode and Effects Analysis (FMEA)

Unlike retrospective RCA, Failure Mode and Effects Analysis (FMEA) is a prospective tool used before implementing a new clinical process or technology (e.g., before launching a new electronic prescribing system or chemotherapy infusion protocol):

• The team maps every step in the proposed workflow and identifies potential failure modes.
• For each failure mode, three scores (from 1 to 10) are assigned:
  1. Severity (S): Seriousness of the clinical effect if the failure occurs.
  2. Occurrence (O): Likelihood that the failure will occur.
  3. Detection (D): Likelihood that the failure will be intercepted before reaching the patient.
• Risk Priority Number (RPN): Calculated as $\text{RPN} = S \times O \times D$ (range 1 to 1,000). Teams prioritize high-RPN failure modes for redesign.

---

Culture of Safety & The "Just Culture" Framework

A Culture of Safety exists when healthcare workers share common perceptions regarding the importance of safety and pursue proactive error reduction without fear of retribution. Key elements include Psychological Safety (the belief that one will not be punished or humiliated for speaking up with ideas, questions, concerns, or mistakes) and robust Near-Miss Reporting:

• Near-Miss (Close Call): An event or situation that could have resulted in accident, injury, or illness, but did not either by chance or through timely interception (e.g., a pharmacist catches an inappropriate pediatric ten-fold antibiotic overdose before the medication is dispensed).
• Significance: Near-misses occur 10 to 100 times more frequently than adverse events. Treating near-misses as free learning opportunities allows organizations to fix latent system flaws before a patient is injured.

The Just Culture Triage Model (David Marx)

Historically, organizations oscillated between a punitive culture (blaming and punishing clinicians for any error, which drives mistakes underground) and a blame-free culture (refusing to hold anyone accountable, even for reckless conduct). A Just Culture creates a balanced, transparent framework that categorizes human behavior into three distinct classes:

---

Communication and Adverse Event Disclosure (CANDOR)

When an adverse medical event occurs, the historical practice was "deny and defend"—withholding information, avoiding contact with families, and defending cases in protracted litigation. Evidence demonstrates that this approach causes severe psychological trauma to patients and clinicians ("second victims"), erodes public trust, and increases malpractice payouts.

The AHRQ CANDOR Process

The Agency for Healthcare Research and Quality (AHRQ) established the CANDOR (Communication and Optimal Resolution) process, endorsed as the gold-standard ethical and operational framework for managing adverse events:

1. Immediate Clinical Care & Safety: Attend to the patient's immediate medical needs, stabilize clinical condition, and secure involved medications and equipment.
2. Transparent Disclosure Within 24 Hours: Hold an empathetic, face-to-face conference with the patient and family. Disclose known objective clinical facts in plain language without premature speculation regarding legal liability or assigning individual fault.
3. Sincere Expression of Empathy: Offer a genuine, compassionate apology expressing regret for the outcome (e.g., "I am deeply sorry that this complication occurred"). In most jurisdictions, expressions of sympathy are protected under state "Apology Laws" and cannot be used as an admission of legal negligence.
4. Commitment to Investigation: Reassure the patient and family that a thorough, multidisciplinary systems investigation is underway, and commit to sharing the findings openly with them.
5. Second Victim Support: Provide immediate psychological, emotional, and peer support to the involved clinical team members.
6. Resolution & Fair Compensation: When investigation reveals that patient harm resulted from medical error, proactively offer fair financial compensation and waive all associated hospital/clinical bills without forcing protracted litigation.

---

Diagnostic Stewardship & Mitigating Low-Value Care

Diagnostic Stewardship Principles

Diagnostic Stewardship is the systematic effort to guide the appropriate selection, collection, and interpretation of diagnostic tests to optimize patient care and reduce the unintended harms of testing. In outpatient primary care, unselected testing frequently triggers the Diagnostic Cascade:

• A clinician orders a low-value screening test in an asymptomatic individual.
• The test detects an ambiguous, clinically irrelevant incidental finding ("incidentaloma").
• The finding triggers anxiety, repeat laboratory tests, specialized CT/MRI imaging, and invasive biopsies.
• The cascade leads to physical procedural complications, substantial out-of-pocket financial toxicity, and clinical harm without improving longevity or quality of life.

Master Table of AAFP Choosing Wisely Recommendations in Family Medicine

---

Medication Reconciliation & Care Transition Safety

The Vulnerability of Care Transitions

Transitions of care—particularly hospital discharge to home, skilled nursing facility (SNF) placement, or emergency department discharge—represent the highest-risk interval for medical errors in primary care. Up to 50% of adult patients experience at least one medication discrepancy at hospital discharge, and 20% experience an adverse drug event (ADE) within 30 days of leaving the hospital.

The 5-Step Medication Reconciliation Process

Medication reconciliation is the formal, systematic process of creating the most comprehensive, accurate list of all medications a patient is currently taking (including prescription drugs, OTC medications, vitamins, and herbal supplements) and comparing it against orders across transitions:

1. Verification: Compile a complete, verified list of the patient's pre-admission medications, cross-referencing patient report, pill bottles, pharmacy dispensing records, and outpatient EHR notes.
2. Clarification: Evaluate appropriateness of dosages, routes, indications, and compliance history.
3. Reconciliation: Compare the verified pre-admission list directly against the proposed transfer or discharge orders. Identify and resolve all discrepancies:
   • Unintended omissions: Chronic home medications inadvertently not resumed at discharge (e.g., home beta-blocker or statin omitted).
   • Unintended duplications: Prescribing both a hospital formulary equivalent and a home agent (e.g., discharging a patient on both pantoprazole and omeprazole).
   • Dosing discrepancies: Inadvertently discharging on an altered acute dose (e.g., high-dose IV steroid converted to excessive outpatient oral maintenance).
4. Documentation: Formulate a reconciled, clean medication list with clear clinical rationale for all additions, discontinuations, and dosage titrations.
5. Transmission & Patient Education: Deliver the reconciled list in clear, understandable language to the patient/caregiver, outpatient primary care physician, and community dispensing pharmacy.

High-Alert Medication Classes in Primary Care Transitions

Adverse events at care transitions disproportionately involve five high-alert medication classes:

• Anticoagulants (Warfarin, DOACs): Dosing confusion, drug interactions, failure to arrange prompt INR monitoring, or co-prescribing with antiplatelet agents.
• Insulins & Hypoglycemic Agents: Failure to adjust hospital sliding scales back to home basal-bolus regimens; acute renal improvement leading to severe sulfonylurea-induced hypoglycemia.
• Opioid Analgesics: Discharging opioid-naive patients on excessive quantities or failing to co-prescribe bowel regimens.
• Cardiovascular & Antihypertensive Agents: Abrupt cessation of clonidine or beta-blockers causing rebound hypertensive crises; combining ACE inhibitors, ARBs, and potassium-sparing diuretics causing lethal hyperkalemia.
• Antimicrobials: Discharging on prolonged empiric courses without tailoring to finalized outpatient microbiology cultures.