{
  "id": "quality-of-care",
  "title": "Quality of Care",
  "lang": "en",
  "source_lang": "en",
  "summary": "Quality of care describes how well health services improve desired health outcomes while meeting standards of effectiveness, safety, responsiveness, and equity.",
  "category": "health",
  "subcategory": "public-health",
  "type": "concept",
  "aliases": [
    "Healthcare quality",
    "Health care quality",
    "Quality of healthcare",
    "Quality of health care"
  ],
  "generated_at": "2026-11-03",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/quality-of-care/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/quality-of-care/",
      "json": "https://aiwiki.page/zh/quality-of-care.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "Quality of care is the degree to which health services for individuals and populations increase the likelihood of desired health outcomes and are consistent with current professional knowledge. It concerns not only whether a treatment works under ideal conditions, but also whether appropriate care reaches the right person, is delivered safely and competently, respects that person’s circumstances and preferences, and remains coordinated over time. Quality applies across prevention, diagnosis, treatment, rehabilitation, and palliative care, and across settings ranging from community services to highly specialized hospitals.\n\nQuality is a central property of a health system, distinct from—but closely connected to—access, expenditure, and financial protection. A service may be available yet ineffective, technically proficient yet disrespectful, or inexpensive yet unsafe. Conversely, adverse outcomes do not necessarily demonstrate poor care, because disease severity, biological uncertainty, and circumstances outside health services also influence results. Assessing quality therefore requires evidence about clinical processes, organizational conditions, outcomes, and people’s experiences, rather than reliance on a single statistic."
    },
    {
      "heading": "Intellectual foundations and historical development",
      "text": "Efforts to assess care long preceded modern quality-management systems. In the nineteenth century, Florence Nightingale used hospital and military health statistics to examine mortality and the conditions surrounding care. The work of Ignaz Semmelweis on hand hygiene and Joseph Lister on antiseptic surgery demonstrated that clinical outcomes could depend profoundly on how care was organized and delivered, rather than solely on the underlying illness. These developments helped establish the idea that variation in outcomes could reveal preventable deficiencies.\n\nIn the early twentieth century, the American surgeon Ernest Amory Codman promoted an “end-result” system: hospitals should follow patients after treatment, determine whether care had succeeded, and investigate failures. His approach challenged the assumption that professional reputation or institutional prestige was sufficient evidence of quality. Hospital standardization subsequently placed increasing emphasis on organized medical staff, accurate records, and defined institutional responsibilities. These efforts contributed to the growth of external inspection and healthcare accreditation, although compliance with organizational standards was never equivalent to proof of good outcomes.\n\nA major conceptual advance came from Avedis Donabedian, whose 1966 article “Evaluating the Quality of Medical Care” distinguished structure, process, and outcome. Structure comprises the resources and arrangements through which care is delivered, including staff, facilities, equipment, and organizational systems. Process concerns what providers and patients actually do during care. Outcome concerns resulting changes in health, functioning, knowledge, or satisfaction. The framework remains influential because it links organizational capacity to clinical activity and then to consequences, while recognizing that these relationships require evidence rather than assumption.\n\nFrom the late twentieth century, quality assessment increasingly incorporated evidence-based medicine, industrial approaches to process improvement, and systematic attention to preventable harm. The US Institute of Medicine’s To Err Is Human, released in 1999, brought patient safety into public and policy debate. Its 2001 report Crossing the Quality Chasm articulated six aims: safe, effective, patient-centered, timely, efficient, and equitable care. These aims shifted attention from isolated professional performance toward the design of care systems.\n\nInternational frameworks subsequently emphasized that quality must accompany expansion of coverage. The World Health Organization describes quality services as effective, safe, people-centered, timely, equitable, integrated, and efficient. The inclusion of integration reflects the growing importance of care across multiple providers and settings. The 2018 joint report by WHO, the Organisation for Economic Co-operation and Development, and the World Bank treated quality as essential to achieving universal health coverage, rather than as an optional refinement after access had been secured."
    },
    {
      "heading": "Dimensions of quality and the meaning of appropriate care",
      "text": "Effectiveness concerns whether care produces benefits supported by reliable knowledge. It differs from efficacy, which generally refers to performance under controlled research conditions. A treatment shown to work in a randomized controlled trial may deliver less benefit in routine practice if diagnosis is delayed, essential supplies are unavailable, or follow-up breaks down. Quality therefore includes the reliable implementation of effective interventions, not merely their presence in guidelines or formularies.\n\nAppropriateness links evidence to the circumstances of a particular person. The same intervention may be beneficial in one clinical situation, unnecessary in another, and harmful in a third. Three recurring failures are underuse, overuse, and misuse. Underuse occurs when beneficial care is omitted; overuse occurs when services offer little or no net benefit; misuse occurs when an otherwise appropriate service is delivered incorrectly. More testing or treatment is consequently not a dependable indicator of higher quality. Overdiagnosis, for example, can identify abnormalities that would never have caused symptoms while exposing people to further procedures and treatment.\n\nSafety concerns avoiding preventable harm caused by care. Timeliness concerns delays that increase suffering, worsen prognosis, or interrupt necessary treatment. The significance of delay depends on the condition: minutes may matter in an acute stroke, whereas a different timescale applies to monitoring a stable chronic illness. Waiting-time measures therefore require clinically meaningful definitions, including when the clock starts, what counts as completion, and how urgency is classified.\n\nPeople-centeredness addresses dignity, communication, responsiveness, and participation. It does not mean providing every requested intervention regardless of evidence. Rather, it recognizes that technically appropriate treatment can still be poor care if the person does not understand it, cannot use it, or experiences avoidable humiliation. Equity concerns unjust differences in access, processes, experience, and outcomes. Integration concerns whether different services work coherently, especially when people cross organizational boundaries.\n\nEfficiency describes the relationship between resources and meaningful results. Reducing waste may improve quality when it eliminates duplicated investigations or unnecessary administrative work. Cost reduction can also undermine quality if it removes essential staff, equipment, or follow-up capacity. Economic evaluation sometimes expresses health gains through the quality-adjusted life-year, but this measure does not capture every relevant feature of care, including dignity, procedural fairness, or continuity.\n\nThese dimensions can reinforce one another, but they can also conflict. Shorter visits may increase appointment availability while reducing time for explanation. Centralizing complex surgery may concentrate expertise while increasing travel burdens. Quality assessment makes such trade-offs explicit rather than assuming that one indicator can represent the entire service. The appropriate balance depends on clinical evidence, patient preferences, resource constraints, and legitimate public priorities."
    },
    {
      "heading": "Measurement, indicators, and interpretation",
      "text": "Quality measurement translates broad goals into observable quantities. Structural indicators include staffing levels, availability of essential equipment, and functioning referral arrangements. Process indicators assess whether eligible patients received a specified action, such as timely administration of an indicated medicine. Outcome indicators include complications, survival, functioning, and quality of life. Experience measures assess communication, respect, access, and coordination. No category is sufficient alone: resources do not guarantee good practice, recorded processes may not reflect actual care, and outcomes have many determinants.\n\nA process measure requires a clearly defined population, numerator, denominator, and observation period. A simplified measure is:\n\n\\[\\text{Process adherence}= \\frac{\\text{eligible care opportunities in which the action occurred}} {\\text{all eligible care opportunities}}\\times100\\%.\\]\n\nEligibility rules matter greatly. Contraindications, documented refusal, or clinical exceptions may change whether an action was appropriate. Excluding these cases incorrectly can inflate performance; including them indiscriminately can penalize sound clinical judgment. Measures based on administrative data are especially sensitive to coding practices, while record review is more detailed but usually more labor-intensive.\n\nOutcome comparisons frequently require risk adjustment because providers treat different populations. A referral center may have higher observed mortality because it receives patients with more severe illness. Adjustment models estimate expected outcomes using factors such as age, diagnosis, severity, and relevant coexisting conditions. A basic observed-to-expected ratio is:\n\n\\[R=\\frac{O}{E},\\]\n\nwhere \\(O\\) is the observed number of events and \\(E\\) is the number expected under the model. For an adverse outcome, a ratio above 1 indicates more events than expected, but it does not establish the cause. Inadequately measured severity, incomplete follow-up, and differences in coding can all affect the result.\n\nMeasurement also requires attention to reliability, validity, and statistical uncertainty. Rare events may produce unstable rates, especially in small hospitals or low-volume specialties. A provider’s apparent rank can change substantially through chance alone. Confidence intervals, multiyear estimates, and statistical models that account for small samples can reduce misleading interpretation. Comparisons also depend on whether outcomes are attributed to the admitting hospital, the treating team, or the wider network responsible for follow-up.\n\nPatient-reported outcomes extend measurement beyond laboratory values and survival by capturing symptoms, functioning, and health-related well-being directly from patients. They differ from patient-reported experience measures, which ask what happened during care. Changes in a biomarker may be useful but do not always establish improvement in outcomes that matter to patients; this is a central limitation of some surrogate endpoints.\n\nComposite scores combine several indicators, but their meaning depends on weighting rules and the handling of missing data. A high overall score can conceal a serious weakness in one domain. Measurement also changes behavior: poorly designed targets may encourage selective documentation, exclusion of complex patients, or concentration on measured activities at the expense of unmeasured care. Indicators are therefore evidence for investigation and accountability, not self-interpreting verdicts."
    },
    {
      "heading": "Patient safety and the organization of reliable care",
      "text": "Patient safety is a major component of quality, but the terms are not interchangeable. A service may avoid obvious injury while still providing ineffective treatment, poor communication, or unnecessary delays. Conversely, an adverse event can occur despite appropriate care. Safety analysis distinguishes harm associated with healthcare from harm caused by the underlying disease, and then examines whether the healthcare-associated harm was preventable.\n\nSafety problems include medication errors, healthcare-associated infections, wrong-site procedures, failures to recognize deterioration, and diagnostic delays. Their mechanisms often involve several interacting conditions. A medication error, for example, may begin with an ambiguous prescription, pass through a poorly designed electronic interface, escape pharmacy verification, and reach a patient during a rushed handover. Describing the event only as an individual’s mistake misses the conditions that allowed successive safeguards to fail.\n\nThe systems approach examines workload, equipment design, communication, staffing, and organizational priorities. It distinguishes immediate errors from latent conditions that make errors more likely or more consequential. Standardization, reliable identification, medication reconciliation, equipment checks, and structured handovers are examples of mechanisms intended to reduce variation where variation creates risk. In infection prevention, hand hygiene, aseptic technique, device management, environmental cleaning, and surveillance address different points in the chain of transmission. Antimicrobial stewardship connects appropriate prescribing with both individual safety and control of antimicrobial resistance.\n\nIncident reporting can reveal hazards and near misses, but the number of reports is not a direct measure of harm. A service with a more open reporting culture may record more incidents than a less transparent service with worse underlying performance. Voluntary reports are also incomplete and unevenly distributed across event types. Safety assessment therefore combines reporting with record review, infection surveillance, pharmacy data, complaints, and direct observation.\n\nInvestigations commonly use root-cause analysis or related systems-based methods to reconstruct events and identify contributing factors. The term “root cause” can be misleading when an event results from multiple interacting weaknesses rather than one removable defect. Corrective actions also differ in strength: a warning email depends on memory and attention, whereas redesigning incompatible connectors can physically prevent certain misconnections.\n\nA just culture seeks to distinguish inadvertent human error from reckless conduct while preserving accountability. Its purpose is neither automatic punishment nor blanket exemption from responsibility. Staff must be able to raise concerns, and organizations must respond to known hazards. Reliable care also depends on the capacity to detect deterioration, adapt to unexpected conditions, and recover when initial plans fail. Standard procedures and adaptive expertise are complementary rather than competing features of safety."
    },
    {
      "heading": "Quality improvement and evaluation of change",
      "text": "Healthcare quality improvement is the systematic effort to make care processes and outcomes better. It differs from measurement alone: measurement describes performance, while improvement changes the conditions producing that performance. It also differs from inspection, which primarily judges compliance, and from research, whose central purpose is to generate knowledge that extends beyond a particular setting. In practice, these activities often overlap.\n\nAn improvement project usually identifies a specific problem, maps the relevant process, establishes a baseline, and tests a proposed change. For example, delayed communication of important test results may reflect unclear responsibility rather than slow laboratory processing. Mapping the pathway from ordering to result review can reveal whether the failure lies in notification, acknowledgment, patient contact, or follow-up arrangements. The proposed intervention must correspond to the failure mechanism; adding an alert does not solve a problem if no one is assigned to act on it.\n\nThe Plan–Do–Study–Act cycle is a widely used method for testing changes on a limited scale, examining their effects, and adapting them before wider implementation. Other approaches include Lean methods, which examine unnecessary steps and interruptions in workflow, and Six Sigma methods, which emphasize variation and defects. These methods provide organizational tools, not guarantees of effectiveness. Their value depends on the problem, the fidelity of implementation, and the quality of evaluation.\n\nClinical audit compares actual practice with explicit standards, implements changes, and reassesses performance. Audit and feedback provide clinicians or teams with information about their performance relative to a target or comparator. The mechanism is not simply the delivery of a report: effects depend on whether the feedback is credible, timely, actionable, and connected to resources or authority to change practice. Education alone may have limited influence when the main barriers are missing supplies, conflicting incentives, or excessive workload.\n\nTime-series methods are particularly useful because improvement unfolds over time. Run charts display observations sequentially; statistical process-control charts add limits that help distinguish routine variation from signals of process change. A simple before-and-after comparison is vulnerable to seasonal effects, changing patient populations, regression to the mean, and simultaneous initiatives. Stronger evaluations may use interrupted time-series designs, comparison sites, or a cluster randomized trial in which wards or facilities, rather than individual patients, receive an intervention.\n\nEvaluation also includes balancing measures. A project that shortens hospital stays may appear successful while increasing unplanned returns or shifting work to families and community services. Sustainability requires more than an initial improvement: responsibilities, staffing, training, and information systems must support continued performance. Interventions also require adaptation across settings. A pathway developed in a well-resourced teaching hospital may rely on laboratory turnaround times or specialist availability that do not exist elsewhere. Replication therefore involves understanding which components produce the benefit and which can change without undermining it."
    },
    {
      "heading": "Patient experience, decision-making, and continuity",
      "text": "Patient experience concerns what people encounter during care: whether staff listen, explanations are understandable, privacy is respected, pain is addressed, and services are coordinated. Satisfaction is related but conceptually different. Satisfaction reflects the relationship between experience and expectations, so two people receiving similar care may report different satisfaction. Experience questions that ask about specific events are often more interpretable than a single global rating.\n\nCommunication is a clinical mechanism, not merely a courtesy. A person who leaves a consultation without understanding the purpose of a medicine or the next step in assessment may be unable to participate effectively in care. Language barriers, sensory impairment, cognitive difficulties, and limited health literacy can all affect information exchange. Professional interpretation, accessible formats, and confirmation of understanding are organizational features studied as components of quality. Written information alone cannot establish that meaningful communication occurred.\n\nShared decision-making is especially relevant when more than one reasonable option exists and the alternatives differ in outcomes important to patients. Clinical evidence informs the likely benefits and harms, while the patient contributes preferences, priorities, and lived circumstances. This is distinct from simply presenting a menu of choices or transferring responsibility to the patient. It also differs from informed consent, which establishes authorization for an intervention; a signed form does not by itself demonstrate either a good decision process or adequate understanding.\n\nContinuity has several dimensions. Relational continuity involves an ongoing therapeutic relationship. Informational continuity means that relevant knowledge follows the person across encounters. Management continuity means that different providers’ plans are coherent and responsive to changing needs. These dimensions are particularly important in primary care, mental health services, and care for people with multimorbidity, who may otherwise receive conflicting instructions from multiple specialists.\n\nTransitions between hospital, home, rehabilitation, and long-term care create predictable risks. Discharge information may omit medication changes, pending investigations, or responsibility for follow-up. A hospital can perform well on an isolated inpatient measure while leaving the next provider without the information required for safe care. Assessing continuity therefore extends the unit of analysis beyond a single encounter or institution.\n\nQuality near the end of life illustrates why patient priorities cannot be reduced to survival alone. In end-of-life care, symptom relief, communication, dignity, and alignment between treatment and expressed goals are central outcomes. Family and caregiver experience also matters, although it does not replace the patient’s own preferences when those can be expressed. Experience surveys, complaints, interviews, and direct participation in service design provide different kinds of evidence; each captures perspectives that routine clinical records may omit."
    },
    {
      "heading": "Equity, access, and population-level quality",
      "text": "Health equity places the distribution of quality at the center of assessment. An average improvement can coexist with widening differences between socioeconomic groups, geographic areas, or people with different language and disability needs. Equity analysis therefore examines who receives care, who is missing from the data, and whether comparable clinical needs receive comparable responses. Equal treatment is not always equitable treatment, because people may require different resources to achieve a comparable opportunity for benefit.\n\nAccess and quality are closely linked but analytically distinct. A clinic may provide excellent care to those who attend while remaining inaccessible to people who cannot afford transport or leave work. Conversely, expanding attendance without adequate staffing or supplies can increase contact without increasing benefit. Effective coverage addresses this distinction by considering whether people in need receive services of sufficient quality to produce the intended health gain. Counting visits, beds, or procedures alone cannot establish effective coverage.\n\nThe social determinants of health influence both clinical needs and the feasibility of care plans. Housing conditions, income, education, and working arrangements affect exposure to illness and the resources available for recovery. Health services do not control all these factors, but service design can either compound or reduce their consequences. For example, a follow-up system that relies entirely on reliable internet access may systematically miss people with limited connectivity.\n\nMeasuring inequity requires appropriate denominators and sufficiently detailed data. Stratification by geography, deprivation, age, disability, or other relevant characteristics can reveal patterns hidden in aggregate results. Small subgroup numbers, missing demographic information, and inconsistent classification can limit interpretation. Collecting sensitive data also creates responsibilities concerning consent, governance, and data privacy.\n\nRisk adjustment creates a particular tension. Accounting for clinical severity may improve fairness in provider comparisons, but adjusting away the effects of disadvantage can make unequal outcomes appear expected and therefore less visible. Reporting clinically adjusted outcomes alongside results stratified by social circumstances can distinguish different questions: whether providers face different case mixes, whether care differs within those groups, and whether the system produces equitable outcomes.\n\nPopulation-level quality also depends on how services fit together. Primary health care, emergency referral systems, maternal services, chronic disease management, and public health programs are interdependent. Reliable community management may prevent avoidable hospital admissions; effective emergency care depends on recognition, transport, referral communication, and hospital capacity. Some admissions remain unavoidable despite high-quality care, making “avoidable admission” rates signals rather than definitive judgments. Geographic coverage, referral completion, continuity, and outcomes together provide a more informative picture than the performance of individual institutions alone."
    },
    {
      "heading": "Governance, incentives, and digital infrastructure",
      "text": "Quality governance assigns responsibility for setting standards, monitoring performance, responding to failures, and supporting improvement. It operates at several levels: clinical teams, healthcare organizations, professional bodies, purchasers, regulators, and governments. Licensing establishes minimum legal conditions for practice or operation. Accreditation typically assesses compliance with organizational standards through external review. Professional certification evaluates specified competencies. These mechanisms address different questions and cannot substitute for direct evidence about patient outcomes.\n\nClinical guidelines translate evidence into recommendations for defined situations, while local protocols specify operational steps. Their quality depends on evidence appraisal, management of conflicts of interest, transparency, and updating. Guidelines can reduce unwarranted variation, but rigid application may be inappropriate when trial populations differ from the person receiving care. This is especially relevant to older people with several conditions, for whom separate disease-specific pathways can produce incompatible treatment burdens.\n\nPayment arrangements influence what organizations can and are encouraged to do. Fee-for-service commonly rewards activity; capitation provides payment for a defined population; bundled payments cover an episode or set of related services. Pay-for-performance explicitly links payment to selected indicators. Each approach creates possible benefits and risks. A narrowly defined target may improve the measured activity while encouraging neglect of other needs, and outcome-based payment may create incentives to avoid unusually complex patients unless safeguards are present. The effects of health financing therefore depend on design, context, and monitoring rather than the payment label alone.\n\nPublic reporting and organizational dashboards can support transparency, but they require intelligible measures and careful explanation of uncertainty. Raw league tables can exaggerate differences or obscure the role of case mix. Complaints, professional review, regulatory inspections, and patient participation provide complementary forms of accountability. A system that collects performance information without assigning responsibility or enabling action may impose substantial reporting work without improving care.\n\nDigital infrastructure changes both the opportunities and the risks. An electronic health record can make medication histories, results, and care plans more accessible, while computerized prescribing and decision support can identify certain contraindications or interactions. Poor interface design, duplicate records, excessive alerts, and copied-forward errors can create new hazards. Interoperability involves not only transferring data but preserving their meaning, provenance, and relevance across systems.\n\nArtificial intelligence introduces further questions about validation, bias, workflow, and accountability. High predictive accuracy does not by itself demonstrate better care: a prediction must lead to an effective action, reach the appropriate professional, and avoid harmful unintended effects. Evaluation consequently distinguishes technical performance, clinical utility, patient outcomes, and distributional effects. Models may also perform differently after changes in population, clinical practice, or data collection. Procurement, post-deployment monitoring, cybersecurity, and clear responsibility for acting on outputs are therefore part of quality governance, alongside the assessment of the technology itself."
    }
  ],
  "links": [
    "diagnosis",
    "rehabilitation",
    "palliative-care",
    "health-system",
    "ignaz-semmelweis",
    "joseph-lister",
    "evidence-based-medicine",
    "world-health-organization",
    "universal-health-coverage",
    "randomized-controlled-trial",
    "overdiagnosis",
    "stroke",
    "quality-adjusted-life-year",
    "surgery",
    "quality-of-life",
    "patient-reported-outcome",
    "biomarker",
    "surrogate-endpoint",
    "patient-safety",
    "antimicrobial-resistance",
    "cluster-randomized-trial",
    "health-literacy",
    "shared-decision-making",
    "primary-care",
    "mental-health",
    "end-of-life-care",
    "health-equity",
    "social-determinants-of-health",
    "data-privacy",
    "primary-health-care",
    "public-health",
    "health-financing",
    "artificial-intelligence"
  ],
  "sources": [],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}