Quality Improvement

Clinical quality improvement is the part of medicine that treats the delivery of care itself as something to be measured, redesigned, and improved.

14 frameworks, from 1924 to today · Clinical Medicine · Medicine & Health

still a working view ended · each leads to its entry below

Overview

Clinical quality improvement is the part of medicine that treats the delivery of care itself as something to be measured, redesigned, and improved. Its practitioners ask not whether a treatment works in a trial but whether the right treatment reaches the right patient reliably, safely, and without waste—and how to change the systems that determine the answer. What is at stake is the gap between what medicine knows and what medicine does: the diabetic whose blood pressure is never checked, the surgical patient who receives the wrong antibiotic, the hospital where the same operation has wildly different outcomes from the one across town. Quality improvement grew up as a set of methods for closing that gap, borrowed from industry, statistics, and engineering, and gradually fused with clinical epidemiology into a discipline with its own journals, reporting standards, and professional identity.

Statistical roots in industry, 1920s–1950s

The intellectual foundation came from manufacturing. In the 1920s Walter Shewhart, working at Bell Telephone Laboratories, developed the control chart: a plot of a measured quantity over time with statistically derived limits that distinguish ordinary variation from a signal that something has changed123. Shewhart's insight was that variation is inherent in any process, and that reacting to every fluctuation ("tampering") makes things worse, while ignoring genuine shifts makes them worse too3. The control chart gave managers a rule for telling the two apart.

W. Edwards Deming carried these ideas to postwar Japan, where from 1950 he taught statistical quality control to Japanese engineers and managers12. Japanese industry built on Shewhart and Deming to develop company-wide quality management, and by the 1970s Japanese manufacturing had become a benchmark. Deming's later "System of Profound Knowledge" framed organizations as systems whose performance is governed mostly by their design, not by individual effort—a claim with direct implications for medicine, where blaming individual clinicians for errors in badly designed processes was the norm.

Donabedian and the framework for judging quality, 1966

Medicine entered the story through Avedis Donabedian, a physician and researcher at the University of Michigan4. In a 1966 paper he proposed that quality of care be assessed through three linked categories: structure (the resources and organization of care—staffing, facilities, equipment), process (what is actually done to and for the patient), and outcome (the resulting health status)4. The framework's power was that it made quality measurable without requiring that every judgment rest on outcomes, which are often rare, delayed, and confounded. If process is reliably linked to outcome, measuring process is a legitimate proxy. Donabedian spent the following decades elaborating the model, including his influential seven "pillars" of quality and his insistence that quality includes interpersonal care and the patient's own assessment. The structure–process–outcome triad remains the default grammar of the field.

Small-area variation and the evidence of unwarranted difference, 1970s

A second current came from health services research. In the 1970s John Wennberg and colleagues at Dartmouth began mapping rates of common procedures—tonsillectomy, hysterectomy, prostatectomy—across small geographic areas of New England. They found differences far larger than disease burden could explain: rates of the same surgery varied several-fold between neighboring communities. The finding, extended nationally and internationally, became known as small-area variation analysis. Its significance was conceptual: if clinicians in adjacent towns, facing similar populations, did such different things, then practice was being driven by local custom and supply rather than by evidence. This undercut the assumption that quality problems were the exception and reframed variation itself as a target for improvement. The Dartmouth work fed directly into the later movement for shared decision-making and into the "unwarranted variation" language of health policy.

Audit, feedback, and the rise of clinical registries, 1980s–1990s

Alongside these came a simpler, older idea formalized into method: measure what clinicians do, compare it to a standard, and feed the comparison back. Clinical audit—systematic review of care against explicit criteria, with feedback to those delivering it—became widespread in the 1980s and 1990s, particularly in the United Kingdom, where it was built into professional and later regulatory expectations. The evidence base for audit and feedback was assembled much later; systematic reviews found modest but real effects, larger when feedback came from a respected source, was repeated, and included explicit targets.

Clinical registries grew in parallel: databases that collect standardized data on all patients receiving a particular treatment or with a particular condition. The earliest large examples include cardiac surgery registries, which from the 1980s allowed institutions to compare risk-adjusted mortality. Registries made benchmarking possible at scale and turned outcome measurement from a research exercise into a routine clinical activity. They also created the infrastructure for public reporting and pay-for-performance, and the controversies that came with them—risk adjustment, gaming, and the possibility that measurement distorts what it measures.

The Model for Improvement and the spread of rapid-cycle change, 1990s

The methods that most define modern clinical quality improvement were codified by the Institute for Healthcare Improvement (IHI), founded in 1991 by Donald Berwick and colleagues. The IHI's Model for Improvement posed three questions—What are we trying to accomplish? How will we know that a change is an improvement? What changes can we make that will result in improvement?—and paired them with the Plan-Do-Study-Act (PDSA) cycle, a small-scale test of change in which a modification is tried, observed, and adjusted before being spread5. The PDSA cycle descends from Shewhart's specification of the cycle and Deming's teaching of it; the IHI's contribution was to package it for clinical teams as a practical engine for local improvement.

The approach was persuasive because it matched the reality of clinical work: change had to be tested in the messy setting where it would be used, not designed in advance and imposed. Its weakness, acknowledged within the field, was that small local tests rarely produced generalizable knowledge—a tension that would drive the later push for rigorous evaluation and reporting standards.

Patient safety and the systems approach, 1999–2000s

The publication of the US Institute of Medicine's To Err Is Human in 1999, estimating that tens of thousands of deaths each year resulted from medical errors, transformed quality improvement into a public and political cause. Its central argument was the systems approach: errors are consequences of flawed processes and defenses, not primarily of careless individuals, and the remedy is redesign of systems—standardization, simplification, forcing functions, and team training—rather than exhortation or punishment. The idea drew on human factors engineering and on the earlier work of psychologists studying industrial accidents, and it reframed the field's relationship to error: reporting should be encouraged, not penalized.

The systems approach produced concrete tools. Checklists were the most visible: a 2009 study led by Atul Gawande, testing a nineteen-item surgical safety checklist developed with the World Health Organization, reported reductions in complications and deaths across eight cities worldwide. Care bundles—small sets of evidence-based practices that, performed together, produce better outcomes than any alone—were developed for ventilator-associated pneumonia and central-line infections; the central-line bundle work in Michigan, reported in 2006, became a landmark demonstration that a bundle plus a culture of safety could nearly eliminate a common hospital infection. These successes were contested in later years—the mechanisms, the durability, and the size of effects were all debated—but they established the bundle and checklist as standard instruments of the field.

Lean, Six Sigma, and the migration of industrial methods, 2000s

As the field matured, it imported further industrial methodologies. Lean, derived from the Toyota Production System, focused on eliminating waste—waiting, unnecessary motion, overprocessing, defects—and on continuous flow and respect for frontline workers. Six Sigma, developed at Motorola in the 1980s and popularized at General Electric, applied statistical methods to reduce defect rates, with a structured problem-solving sequence (define, measure, analyze, improve, control) and trained practitioners at "belt" levels. Hospitals adopted both, often blended as "Lean Six Sigma," for projects ranging from operating-room turnover to pharmacy inventory to emergency-department flow. The methods brought disciplined measurement and process mapping into clinical settings, but also criticism: that they could be applied mechanically, that their language alienated clinicians, and that evidence of patient-outcome benefit was thinner than their popularity suggested.

Evidence-based guidelines and implementation, 1990s–2010s

A parallel strand concerned getting known-effective care into practice. Clinical practice guidelines, systematized from the 1980s onward and increasingly built on systematic reviews, made explicit what should be done. But guidelines alone changed little; the field of implementation science emerged to study why. Its central finding was that passive dissemination—publication, mailing, lectures—rarely alters behavior, while active strategies such as reminders, audit and feedback, academic detailing, and multifaceted interventions have modest but measurable effects. The recognition that evidence does not implement itself became a defining insight of the field, and it linked quality improvement to the broader evidence-based medicine movement.

Standardization from outside: accreditation and external standards, 1990s–present

External bodies also shaped practice. Accreditation systems—the Joint Commission in the United States, founded in 1951 and evolving through the 1990s into a standards-driven survey process; national programs in the UK, Australia, and elsewhere—set requirements for safety, governance, and quality monitoring, and made participation a condition of payment or licensure in many systems. Accreditation's evidence base has been debated, but its institutional effect was large: it created quality departments, mandated measurement, and made quality a board-level concern. Public reporting of performance and pay-for-performance schemes extended the logic of external accountability.

Standards for reporting improvement work, 2008

As the field accumulated projects, it faced a credibility problem: improvement reports were often too vague to be replicated or judged. In 2008 the SQUIRE guidelines (Standards for QUality Improvement Reporting Excellence) were published, later revised as SQUIRE 2.0 in 2015, specifying what an improvement report should describe—the problem, the intervention, the context, the measures, and the reasoning behind changes. SQUIRE was an attempt to give quality improvement the kind of reporting discipline that clinical trials had acquired decades earlier, and it marked the field's self-conscious turn toward producing generalizable knowledge rather than only local change.

The present

Contemporary clinical quality improvement draws on all of these lineages: the statistical process thinking of Shewhart and Deming, Donabedian's measurement framework, the variation research of Wennberg, the rapid-cycle methods of the IHI, the systems view of patient safety, the industrial toolkits of Lean and Six Sigma, and the accountability machinery of accreditation and public reporting456. Its main approaches today combine local PDSA-style improvement with rigorous evaluation—interrupted time series, stepped-wedge designs, and mixed methods—and with implementation science's attention to context and sustainability. Open questions remain sharp: how to distinguish genuine improvement from measurement artifact and regression to the mean; how to scale successful local projects without losing what made them work; how to balance external accountability against clinician engagement; and how to build a cumulative evidence base for improvement methods themselves, which have often spread faster than the evidence supporting them.

References

  1. Quality management (History), Wikipedia
  2. Statistical process control (History), Wikipedia
  3. Control chart (introduction, History), Wikipedia
  4. Donabedian model (introduction, History), Wikipedia
  5. PDCA (introduction, History and characteristics), Wikipedia
  6. Dartmouth Institute for Health Policy and Clinical Practice, Wikipedia

The frameworks of Quality Improvement

Each framework is a way of working the field has used: when it held, and the question it answers.

  1. 1924-Present

    Statistical Process Control

    How do we tell real change from ordinary variation?

    Fly to its tower in Nousfarer

  2. 1966-Present

    Donabedian model

    How should the quality of care be measured?

    Fly to its tower in Nousfarer

  3. 1973-Present

    Small-Area Variation Analysis

    Why does practice differ so much between places?

    Fly to its tower in Nousfarer

  4. 1980-Present

    Clinical Audit and Feedback

    Does telling clinicians how they perform change it?

    Fly to its tower in Nousfarer

  5. 1980-Present

    Clinical Registries and Outcome Benchmarking

    How do institutions compare their outcomes fairly?

    Fly to its tower in Nousfarer

  6. 1990-Present
  7. 1990-Present
  8. 1991-Present

    Model for Improvement

    How can a team test a change quickly?

    Fly to its tower in Nousfarer

  9. 1998-Present

    Implementation Science

    Why do evidence-based practices spread unevenly?

    Fly to its tower in Nousfarer

  10. 1999-Present

    Systems Approach to Patient Safety

    Why do errors happen in well-meaning organizations?

    Fly to its tower in Nousfarer

  11. 2000-Present

    Six Sigma

    How do we reduce defects to near zero?

    Fly to its tower in Nousfarer

  12. 2001-Present

    Checklists and Care Bundles

    Can simple standard steps prevent harm?

    Fly to its tower in Nousfarer

  13. 2002-Present

    Lean Healthcare

    How do we remove waste from a process?

    Fly to its tower in Nousfarer

  14. 2008-Present

    SQUIRE Reporting Standards

    How should improvement work be reported?

    Fly to its tower in Nousfarer

How the ideas connect

Each arrow runs from an earlier idea to the later one it shaped. The list below says what passed between them.

Statistical Process ControlDonabedian modelSmall-Area Variation AnalysisClinical Audit and FeedbackClinical Registries and Outcome Bench…Accreditation and External Accountabi…Evidence-Based Guideline Implementati…Model for ImprovementImplementation ScienceSystems Approach to Patient SafetySix SigmaChecklists and Care BundlesLean HealthcareSQUIRE Reporting Standards

Beyond the field

Where Quality Improvement's ideas came from in other fields, and what they led to there: the strongest 12 of 59 links.

A city in Nousfarer: a red-carpeted stair climbing terraces to a crowned tower on an acropolis, more towers round it, forest and snowy mountains beyond.
A city in Nousfarer, the game played in the browser.

Quality Improvement is a planet in Nousfarer

Its city has a tower for each of these 14 frameworks. Land, walk in, and climb one concept at a time.

Land on Quality Improvement

Related fields

Fields its ideas are linked to

More in Clinical Medicine

See where Quality Improvement sits on the Map of Human Knowledge