Quality Management, Accreditation and Patient Safety

Introduction

Quality management in hospitals is the organized effort to ensure that care is safe, effective, timely, efficient, equitable, patient-centred and continuously improved. It is not limited to inspection or error counting. It includes the design of reliable systems, measurement of performance, learning from failures, engagement of patients and professionals, and accountability to regulators and the community.

Hospitals are complex socio-technical organizations. Their outcomes depend on interactions among clinicians, patients, equipment, information systems, medicines, buildings, policies and external agencies. Consequently, poor outcomes cannot usually be explained only by the actions of one individual. Modern quality management examines both human performance and the conditions in which work occurs.

Accreditation provides an external framework for assessing whether an organization meets defined standards. Patient safety focuses specifically on preventing avoidable harm and reducing risk when harm cannot be completely eliminated. Clinical governance connects these activities with professional accountability, evidence-based practice and organizational leadership.

Learning Objectives

By the end of this chapter, the learner should be able to:

1. Explain the dimensions and theories of healthcare quality.
2. Distinguish quality assurance, quality improvement, accreditation and clinical governance.
3. Select and interpret appropriate quality indicators and audit methods.
4. Apply incident reporting, root-cause analysis and Failure Mode and Effects Analysis (FMEA).
5. Explain the practical use of Lean and Six Sigma in hospitals.
6. Design patient-safety interventions that address system as well as individual factors.
7. Evaluate accreditation readiness, ethical responsibilities and the limitations of quality data.

1. Meaning and Dimensions of Quality

The Institute of Medicine describes high-quality care through six dimensions: safety, effectiveness, patient-centredness, timeliness, efficiency and equity. These dimensions may conflict. For example, a highly individualized service may require more time and resources, while rapid throughput may create risks if assessment is inadequate. Hospital managers must therefore balance dimensions rather than improve one in isolation.

A useful classification is:

Quality is often examined through structure, process and outcome. Structure includes staffing, facilities, equipment and policies. Process concerns what is done, such as hand hygiene or medication reconciliation. Outcomes include mortality, complications, functional recovery and patient experience. A poor outcome does not automatically prove poor care, because outcomes are influenced by case severity and social factors. Conversely, good outcomes do not prove that processes are reliable.

2. Major Quality-Management Approaches

Quality assurance and quality improvement

Quality assurance is primarily concerned with confirming that predefined requirements are met. It commonly uses policies, credentialing, inspection, audit and corrective action. Quality improvement (QI) is broader and more developmental: teams study processes, test changes, measure results and learn iteratively.

Quality assurance asks, “Are we meeting the required standard?” Quality improvement asks, “How can the system perform better and more reliably?” Both are necessary. Compliance without improvement can produce bureaucratic activity, while improvement without minimum standards can produce unsafe variation.

Donabedian’s model

Avedis Donabedian’s structure–process–outcome model remains a foundational framework. For example, a hospital may assess whether it has sufficient infusion pumps (structure), whether nurses follow an infusion protocol (process), and whether infusion-related adverse events have declined (outcome). The model is not a simple causal formula. It is an organizing framework, and each relationship must be tested in context.

Continuous improvement and the PDSA cycle

The Plan–Do–Study–Act cycle supports small-scale, evidence-informed testing:

• Plan: Define the problem, objective, intervention, measures and predicted result.
• Do: Implement on a limited scale and document what occurred.
• Study: Compare results with the prediction and examine unintended effects.
• Act: Adopt, adapt or abandon the change, then begin another cycle.

PDSA should not be treated as a shortcut for rigorous evaluation. A poorly defined aim, inadequate measurement or premature conclusion can create false confidence. Changes should be tested under the conditions in which they will ultimately operate.

Quality improvement versus research

QI usually seeks to improve a local service using existing knowledge, whereas research seeks generalizable knowledge through a systematic investigation. The distinction can be difficult. Organizations should consult their ethics committee or institutional review board when a project involves experimentation, randomization, identifiable data or publication of generalizable findings.

3. Clinical Governance

Clinical governance is the integrated framework through which health organizations are accountable for continuously improving quality and safeguarding high standards of care. Its elements commonly include:

• evidence-based practice and clinical effectiveness;
• education, training and professional development;
• clinical audit;
• risk management and patient safety;
• incident reporting and learning;
• patient and public involvement;
• information governance and reliable records;
• leadership, accountability and transparency.

Governance is not the responsibility of a quality department alone. The governing board sets expectations, executives allocate resources, clinical leaders guide practice, and frontline teams identify operational risk. Effective governance has clear escalation routes: a serious risk must reach the level able to authorize corrective action.

A mature governance system distinguishes accountability from blame. Professionals remain responsible for competent practice and honest reporting, but organizations must also examine workload, staffing, communication, equipment design, training and policies.

4. Accreditation and Standards

Accreditation is a formal external assessment against published standards by an independent or authorized body. It differs from licensure, which is normally a legal permission to operate, and certification, which may attest conformity to a specified system or competence requirement.

Accreditation standards commonly cover:

1. governance and leadership;
2. patient rights and informed consent;
3. assessment, diagnosis and care planning;
4. medication management;
5. infection prevention and control;
6. surgery and anaesthesia;
7. emergency and critical care;
8. laboratory, imaging and blood services;
9. facility safety and emergency preparedness;
10. health information management;
11. workforce competence;
12. quality improvement and patient safety.

Accreditation readiness

Readiness should be continuous rather than a short-term preparation exercise. A practical approach is to create a standards matrix containing the requirement, responsible owner, evidence, current gap, corrective action, due date and verification method. Evidence should demonstrate actual practice, not merely the existence of a policy. Surveyors may examine records, interview staff and patients, observe care and trace a patient journey across departments.

Common accreditation failures include outdated policies, incomplete documentation, inconsistent practice between shifts, weak credentialing, inadequate medication reconciliation and lack of evidence that corrective actions were sustained. “Mock surveys” are useful only when they test real systems rather than rehearse staff responses.

Accreditation can improve standardization and public trust, but it has limitations. Organizations may focus on visible compliance, burden staff with documentation or neglect unmeasured aspects of care. Standards should therefore support—not replace—professional judgment, patient involvement and outcome measurement.

5. Measurement, Indicators and Audit

A quality indicator is a defined, measurable variable used to monitor performance. Indicators may be classified as:

• Structure indicators: resources or organizational capacity.
• Process indicators: whether recommended actions occurred.
• Outcome indicators: results experienced by patients.
• Balancing indicators: unintended consequences elsewhere in the system.

Examples include hand-hygiene compliance, time to antibiotic administration, pressure-injury incidence, readmission rate and staff overtime. A good indicator has a clear numerator, denominator, population, time period, data source, owner and interpretation rule.

For a proportion or rate:

The multiplier may be 100 for a percentage, 1,000 for events per 1,000 patient-days or another stated value. The numerator must be included in the denominator definition. For example, a catheter-associated infection rate may use infections divided by catheter-days, multiplied by 1,000. Rates based on different denominators must not be compared casually.

Worked example: interpreting an indicator

A ward records 12 falls during 2,400 occupied bed-days. The fall rate per 1,000 occupied bed-days is:

This calculation assumes that all falls were detected and that occupied bed-days were counted consistently. The rate alone does not establish whether care was unsafe. Managers should examine patient acuity, reporting completeness, location, time, injury severity and preventive practices.

Audit

Clinical audit is a systematic comparison of current practice with explicit criteria, followed by action and remeasurement. A complete audit cycle includes selecting a topic, defining standards, collecting baseline data, analyzing gaps, implementing change, repeating measurement and reporting results. An audit that ends with data collection is not a completed improvement cycle.

Sampling may be used when reviewing every record is impractical. The sample should be representative and the method documented. Inconsistent case definitions, missing records and selective data collection can produce misleading conclusions. Confidentiality must be maintained, and data access should follow applicable law and organizational policy.

6. Patient-Safety Frameworks

Patient safety is the prevention of errors and adverse effects associated with healthcare. Important concepts include:

• Adverse event: an incident causing harm.
• Near miss: an incident that could have caused harm but did not, either by chance or timely intervention.
• Never event: a serious, clearly identifiable and preventable event defined by the relevant authority.
• Hazard: a circumstance with potential to cause harm.
• Risk: the possibility of harm combined with its potential severity.
• Safety culture: shared values and behaviours concerning risk, learning, communication and accountability.

James Reason’s “Swiss cheese” model explains how harm may occur when weaknesses in several defensive layers align. Layers may include policies, training, alarms, supervision and physical barriers. The model emphasizes that accidents usually arise from interacting failures rather than a single mistake.

A just culture seeks a fair response to human error, at-risk behaviour and reckless behaviour. Human error generally requires consolation, system redesign and learning. At-risk behaviour requires coaching and removal of incentives that encourage unsafe shortcuts. Reckless behaviour—conscious disregard of substantial risk—may require disciplinary action. A just culture is not a “no-accountability” culture.

High-reliability organizations maintain preoccupation with failure, reluctance to simplify, sensitivity to operations, commitment to resilience and deference to expertise. Hospitals cannot eliminate uncertainty, but they can build systems that detect weak signals and recover quickly.

7. Incident Reporting and Learning Systems

An incident-reporting system should be accessible, non-punitive for good-faith reporting, confidential where appropriate and linked to timely feedback. Reports should describe what happened, where and when it occurred, who was affected, immediate actions, contributing conditions and recommended escalation. Reporting numbers should not be interpreted alone: a rise may indicate worsening safety, improved detection or both.

A useful response sequence is:

1. Protect the patient and provide immediate clinical care.
2. Escalate according to severity and legal requirements.
3. Preserve relevant records and equipment.
4. Communicate honestly and compassionately with the patient and family.
5. Support involved staff, including second victims.
6. Investigate proportionately.
7. Implement, assign and monitor corrective actions.
8. Share learning while protecting confidentiality.

Disclosure should be truthful, timely and respectful. Concealing an error violates professional duties and undermines trust. Local laws determine reporting obligations, privilege and documentation requirements.

8. Root-Cause Analysis

Root-cause analysis (RCA) is a structured investigation of how and why an incident occurred, with the aim of reducing recurrence. It should identify system vulnerabilities rather than seek a convenient individual to blame.

A sound RCA normally includes a multidisciplinary team, a clear event definition, a chronology or timeline, review of policies and records, interviews with involved staff, identification of contributing factors and development of stronger actions. The “five whys” can help explore causal chains, but it should not be used mechanically. Asking “why?” five times does not guarantee discovery of a root cause.

Contributing-factor categories may include patient factors, task design, technology, individual performance, team communication, work environment, organizational policies and external pressures. The investigation should distinguish direct causes, latent conditions and factors that merely coexisted without contributing.

Corrective actions should follow a hierarchy of effectiveness:

• strong actions: physical or forcing functions, automation with independent verification, standardization and elimination of unnecessary steps;
• intermediate actions: checklists, reminders, training and double checks;
• weak actions: warnings, memoranda and reliance on vigilance.

Training alone is usually weak because it does not remove hazards or compensate for poor system design. Every action should have an owner, deadline, measure and follow-up review.

9. Failure Mode and Effects Analysis

FMEA is a prospective method used before harm occurs. A team lists process steps, identifies possible failure modes, estimates their effects and causes, and designs controls. Traditional FMEA may assign ratings for severity, occurrence and detectability. A risk priority number (RPN) is sometimes calculated as:

where S is severity, O is occurrence and D is detectability, each scored according to the organization’s defined scale. The product is dimensionless and is meaningful only within that scoring system. RPN should not be treated as a precise probability, and a low RPN does not automatically mean low risk. High-severity failures may require action regardless of the product.

For example, in medication administration, a failure mode may be selection of a look-alike medicine. Effects may include severe patient harm; causes may include similar packaging and storage proximity; controls may include segregation, barcode verification and independent review for high-alert medicines. The team should evaluate whether the proposed controls work in actual practice.

10. Lean and Six Sigma in Hospitals

Lean focuses on value from the patient’s perspective and reduction of waste. Common forms of waste include defects, overproduction, waiting, unused talent, transportation, inventory, motion and overprocessing. Value-stream mapping can reveal delays across registration, assessment, diagnostics, treatment and discharge.

Lean must not mean simply working faster or reducing staff. Removing waste without protecting safety can increase risk. Patient value, clinical appropriateness and staff wellbeing must remain central.

Six Sigma emphasizes reduction of process variation and defects through the DMAIC cycle: Define, Measure, Analyze, Improve and Control. It is useful when a problem is measurable and a stable process can be studied. Statistical tools may include run charts, control charts, Pareto analysis and process capability analysis. Sophisticated statistics are not a substitute for accurate definitions and trustworthy data.

A basic defect proportion is:

If 18 of 600 laboratory specimens are rejected, the rejection rate is 3%. Interpretation requires knowledge of specimen type, collection conditions and whether the denominator represents all eligible specimens.

Lean and Six Sigma are complementary: Lean emphasizes flow and waste, while Six Sigma emphasizes variation and defects. Both require frontline participation and leadership support.

11. Managing Risk in Practice

Risk management includes risk identification, assessment, control, monitoring and communication. A risk register may record the hazard, affected population, existing controls, likelihood, consequence, risk rating, owner, action and review date. Scoring systems should be calibrated; a numerical score is a decision aid, not an objective truth.

Controls should follow the hierarchy of controls: eliminate the hazard, substitute a safer alternative, apply engineering controls, establish administrative controls and use personal protective equipment. In hospitals, examples include eliminating unnecessary manual transcription, using smart infusion pumps, standardizing protocols and providing protective equipment.

Patient safety should be integrated into procurement, construction, staffing, digital-system design and emergency planning. A new electronic prescribing system, for example, may reduce illegible orders but create risks from alert fatigue, incorrect default doses or downtime. Safety testing must include usability, workflow, interoperability, cybersecurity and contingency procedures.

12. Digital Health and Artificial Intelligence

Electronic health records, barcode medication administration, clinical decision support, dashboards and telehealth can strengthen quality, but technology can also introduce new hazards. Organizations should monitor wrong-patient selection, copy-forward documentation, interoperability failures, downtime and excessive alerts.

AI tools require validation in the population and setting where they will be used. Managers should examine performance across demographic groups, data drift, explainability, human oversight, privacy, cybersecurity and liability. An algorithmic recommendation must not replace professional judgment or informed patient participation. AI-generated documentation should be checked by a competent clinician before becoming part of the legal health record.

13. Illustrative Case: Wrong-Patient Medication Administration

Background

A medical ward reports two near misses in which medicines were almost administered to patients with similar names. The electronic medication system is available, but staff frequently work under time pressure during the evening medication round.

Problem

The initial response recommends retraining nurses to “be more careful.” Incident review shows that two patients share similar names, wristbands are partly obscured by bedding, the medication cart is positioned near a noisy family waiting area, and barcode scanners intermittently fail. The staffing pattern leaves one nurse responsible for a large number of patients.

Analysis

An RCA team constructs a timeline and interviews nurses, pharmacists, patients and information-technology staff. Contributing factors include similar-name alerts that are frequently ignored, storage of medicines in alphabetical order, scanner battery failures, interruptions and an expectation that medication rounds should be completed within a fixed time. No single nurse’s action explains the risk.

A prospective FMEA rates wrong-patient selection as a high-severity failure mode. The team identifies stronger controls: a two-identifier requirement, visual differentiation of similar names, reliable barcode equipment, separated storage of look-alike medicines, protected medication-round zones and a process for reporting scanner failures.

Approach and outcome

The ward pilots the changes for four weeks. Measures include barcode scanning compliance, observed use of two identifiers, near-miss reports, medication-round duration and staff-reported interruptions. Barcode compliance improves from 82% to 97%, and near misses decline. Medication-round duration increases slightly at first, then returns near baseline after equipment and workflow problems are corrected. The ward continues monthly monitoring because early success does not prove sustained reliability.

Learning points

1. The visible error may be the endpoint of multiple system weaknesses.
2. Near misses are valuable signals, not evidence that the system is safe.
3. Stronger design controls are preferable to reminders alone.
4. Balancing measures are needed to identify unintended delays or workarounds.
5. Staff and patients should participate in the redesign.

14. Challenges, Ethics and Best Practices

Quality data can be distorted by underreporting, inconsistent definitions, risk-adjustment problems, missing data, gaming and excessive focus on measured targets. Public reporting may improve accountability but can also encourage avoidance of high-risk patients if adjustment is inadequate. Managers should present uncertainty, limitations and trends rather than isolated rankings.

Ethically, quality management must respect autonomy, confidentiality, justice, non-maleficence and beneficence. Patient involvement should be meaningful, including participation in incident reviews and service design when appropriate. Confidential reports must be protected, while transparency about harm must be maintained.

Best practices include:

• align quality goals with organizational strategy and patient priorities;
• use a small number of clearly defined indicators;
• combine outcome, process and balancing measures;
• involve multidisciplinary frontline teams;
• use data for learning rather than punishment;
• test changes before large-scale implementation;
• assign owners and deadlines for corrective actions;
• verify sustainability after initial improvement;
• communicate openly with patients and families;
• maintain accreditation readiness continuously;
• review digital tools for new forms of risk;
• support staff wellbeing and psychological safety.

Chapter Summary

Quality management is a system for designing, measuring and improving healthcare rather than merely inspecting completed work. Quality includes safety, effectiveness, patient-centredness, timeliness, efficiency and equity. Donabedian’s structure–process–outcome model, PDSA, clinical governance, audit and accreditation provide complementary ways to organize improvement.

Accreditation assesses conformity with defined standards, but meaningful quality requires evidence that safe practices are implemented and sustained. Indicators must have clear definitions and denominators, while audit requires remeasurement after corrective action. Incident reporting, just culture and the Swiss cheese model support learning from harm and near misses. RCA is retrospective; FMEA is prospective. Lean reduces waste and improves flow, while Six Sigma reduces variation and defects. Neither should compromise clinical safety.

Effective patient safety depends on system design, strong leadership, reliable data, patient engagement, ethical transparency and continuous learning. Technology and AI can support quality only when validated, monitored and governed responsibly.

Key Terms

Quality management: Coordinated activities used to direct and control an organization with regard to quality.

Quality assurance: Activities that verify whether predefined requirements are being met.

Quality improvement: Systematic, continuous efforts to improve processes and outcomes.

Accreditation: External assessment against defined healthcare standards.

Clinical governance: Framework for accountability, quality improvement and safe clinical practice.

Clinical audit: Comparison of practice with explicit criteria followed by action and remeasurement.

Quality indicator: A measurable variable used to monitor performance.

Adverse event: An incident that causes harm to a patient.

Near miss: An incident that could have caused harm but did not.

Safety culture: Shared values and behaviours concerning risk, reporting and learning.

Just culture: Fair treatment that distinguishes human error, at-risk behaviour and reckless behaviour.

Root-cause analysis: Retrospective investigation of contributing system factors.

FMEA: Prospective analysis of possible process failures and their effects.

Risk priority number: Product of severity, occurrence and detectability scores in traditional FMEA.

Lean: Improvement approach focused on patient value and waste reduction.

Six Sigma: Method for reducing process variation and defects.

Balancing measure: Measure used to detect unintended consequences of an intervention.

Patient-centred care: Care that respects patient preferences, needs, values and dignity.

High-reliability organization: Organization that operates safely despite complexity and uncertainty.

Review Questions

Short Answer Questions

1. Define quality improvement and distinguish it from quality assurance.
2. List the six dimensions of healthcare quality.
3. What is the difference between a near miss and an adverse event?
4. State the main stages of the PDSA cycle.
5. Why are balancing measures important in quality improvement?

Long Answer Questions

1. Explain Donabedian’s structure–process–outcome model using a hospital example.
2. Discuss the role of accreditation in hospital governance and identify its limitations.
3. Explain how incident reporting and just culture support patient safety.
4. Compare root-cause analysis with FMEA in terms of purpose, timing and method.
5. Describe how Lean and Six Sigma can be combined in a hospital improvement project.

Analytical / Application Questions

1. A ward records 15 medication errors during 3,000 medication administrations. Calculate the error percentage and explain why the result should not be interpreted without a case definition and review of severity.
2. An increase in incident reports occurs after a non-punitive reporting policy is introduced. Analyze at least three possible interpretations and recommend additional data needed.
3. Design a quality-improvement plan for reducing delays in emergency-discharge processing. Include an aim statement, process and outcome indicators, a balancing measure, one PDSA test and a sustainability plan.

Multiple Choice Questions with Answers

Question 1

Which dimension of quality asks whether care achieves its intended clinical results?

A. Efficiency

B. Effectiveness

C. Equity

D. Timeliness

Correct Answer: B. Effectiveness

Question 2

Which activity is prospective rather than retrospective?

A. Root-cause analysis after an adverse event

B. Review of a completed complaint

C. Failure Mode and Effects Analysis

D. Investigation of a medication incident

Correct Answer: C. Failure Mode and Effects Analysis

Question 3

A ward records 12 falls during 2,400 occupied bed-days. What is the fall rate per 1,000 occupied bed-days?

A. 0.5

B. 2

C. 5

D. 20

Correct Answer: C. 5

Question 4

Which is the strongest corrective action for preventing selection of the wrong medicine?

A. Send a reminder email

B. Provide annual retraining only

C. Place a warning poster near the cupboard

D. Use barcode verification with segregated storage

Correct Answer: D. Use barcode verification with segregated storage

Question 5

In Donabedian’s model, hand-hygiene compliance is primarily a:

A. Structure measure

B. Process measure

C. Outcome measure

D. Balancing measure

Correct Answer: B. Process measure

Question 6

The primary purpose of a balancing measure is to:

A. Replace all outcome measures

B. Detect unintended effects of an intervention

C. Determine staff salaries

D. Prove that accreditation has been achieved

Correct Answer: B. Detect unintended effects of an intervention

Question 7

In traditional FMEA, the risk priority number is calculated as:

A. Severity + occurrence + detectability

B. Severity × occurrence × detectability

C. Occurrence ÷ severity × detectability

D. Detectability − occurrence − severity

Correct Answer: B. Severity × occurrence × detectability

Question 8

Which response is most consistent with a just culture after an unintentional human error?

A. Automatic dismissal

B. Public identification of the staff member

C. System learning, support and redesign

D. Ignoring the event because no harm occurred

Correct Answer: C. System learning, support and redesign

Question 9

Which sequence correctly represents DMAIC?

A. Define, Measure, Analyze, Improve, Control

B. Design, Monitor, Audit, Inspect, Correct

C. Describe, Measure, Assess, Integrate, Certify

D. Define, Manage, Approve, Implement, Communicate

Correct Answer: A. Define, Measure, Analyze, Improve, Control

Question 10

Which statement about accreditation is most accurate?

A. It is identical to professional licensure

B. It guarantees that no adverse events will occur

C. It assesses conformity with defined standards

D. It eliminates the need for clinical judgment

Correct Answer: C. It assesses conformity with defined standards