{
  "id": "patient-reported-outcome",
  "title": "Patient-Reported Outcome",
  "lang": "en",
  "source_lang": "en",
  "summary": "A patient-reported outcome is information about health status provided directly by a patient, without interpretation by a clinician or another observer.",
  "category": "health",
  "subcategory": "medicine",
  "type": "concept",
  "aliases": [
    "PRO",
    "Patient-reported health outcome"
  ],
  "generated_at": "2026-11-04",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/patient-reported-outcome/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/patient-reported-outcome/",
      "json": "https://aiwiki.page/zh/patient-reported-outcome.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "A patient-reported outcome (PRO) is a report about a person’s health condition that comes directly from that person, without interpretation by a clinician or anyone else. It may describe symptoms, physical functioning, emotional well-being, or the effects of illness and treatment on daily life. PROs complement clinical examinations, laboratory tests, and other measurements by documenting aspects of health that patients experience themselves. The questionnaire, scale, or other instrument used to collect this information is called a patient-reported outcome measure (PROM)."
    },
    {
      "heading": "Definition and scope",
      "text": "The defining feature of a PRO is the source of the information, not whether the information is numerical or collected electronically. A patient may complete a questionnaire independently or answer questions administered by an interviewer. Interviewer administration remains patient-reported when the interviewer records the patient’s responses without interpreting them.\n\nCommon domains include:\n\n- Symptoms: the intensity, frequency, or interference of pain, fatigue, nausea, breathlessness, and other experiences.\n- Physical functioning: perceived ability to walk, dress, perform household tasks, or participate in work.\n- Emotional and social functioning: emotional distress, relationships, and participation in social roles.\n- Health-related quality of life: the effects of health on physical, psychological, and social well-being.\n- Treatment-related effects: symptoms and limitations experienced during or after treatment.\n\nA PRO is not necessarily a change over time. A report of current pain severity is a PRO; comparing that report with an earlier assessment can describe improvement or deterioration.\n\nRelated but distinct concepts\n\nPatient-reported outcome and patient-reported outcome measure are often used interchangeably, but the outcome is the concept being assessed, whereas the measure is the instrument used to assess it. A PROM generally includes questions or tasks, instructions, response options, a recall period, and scoring rules.\n\nHealth-related quality of life is one possible subject of a PRO, not a synonym for all PROs. A single symptom rating can be a PRO without measuring quality of life.\n\nPatient-reported experience measures assess experiences of care, such as whether explanations were understandable or appointments were accessible. These differ from health outcomes. Treatment satisfaction and treatment preferences are also related forms of patient input but do not automatically constitute PROs about health status.\n\nReports by clinicians, observations by caregivers, and performance tests are other types of clinical outcome assessment. A caregiver’s account of a patient’s condition is not a direct patient report. Observable behavior and inferred internal experience also require different interpretation: a caregiver can observe vomiting but cannot directly experience another person’s nausea."
    },
    {
      "heading": "Types of measures",
      "text": "Generic and condition-specific measures\n\nGeneric measures assess health concepts relevant across different conditions and populations. They support comparisons between diseases, services, or population groups. Examples include the SF-36, EQ-5D, and measures from the Patient-Reported Outcomes Measurement Information System, commonly called PROMIS.\n\nCondition-specific measures focus on problems relevant to a particular disease or clinical context. They may capture effects that a broad generic measure overlooks, but their scores are often less suitable for comparisons across unrelated conditions.\n\nMeasures may also be domain-specific, assessing a concept such as fatigue or physical function across several diseases. Generic, condition-specific, and domain-specific approaches can be used together.\n\nSingle-item and multi-item measures\n\nA single-item measure may ask a patient to rate pain on a numerical scale. Multi-item instruments use several questions to assess a broader or more complex concept. Some provide separate domain scores, while others also produce an overall score.\n\nA total score is not automatically meaningful simply because responses can be added together. Its interpretation depends on the instrument’s design and measurement evidence.\n\nPreference-based measures\n\nSome instruments describe health states that can be assigned preference-based values. These values may contribute to quality-adjusted life-year calculations in cost-effectiveness analysis.\n\nThe patient-reported health description and its assigned value are distinct. A value set may reflect preferences elicited from a general population or another specified group rather than the individual respondent’s own valuation."
    },
    {
      "heading": "Development and measurement properties",
      "text": "Developing a PROM involves specifying what it should measure, whom it is intended for, and how its scores will be used. Interviews and qualitative research with patients help establish whether questions address relevant experiences and are understood as intended.\n\nSeveral measurement properties are important:\n\nProperty | Main question\nContent validity | Are the questions relevant, comprehensive, and understandable for the intended population and purpose?\nReliability | Are scores sufficiently consistent when the underlying health concept has not changed?\nConstruct validity | Do scores behave as expected in relation to other measures and differences between groups?\nResponsiveness | Can the measure detect changes in the concept it is intended to assess?\nMeasurement error | How much variation reflects measurement rather than real differences or change?\n\nInternal consistency concerns relationships among items intended to measure a common construct. It does not, by itself, establish that an instrument measures the right concept or that its scores are valid.\n\nMany patient experiences have no definitive external standard. Validation therefore often relies on multiple forms of evidence rather than comparison with a single “gold standard.” Evidence supporting use in one population, language, or setting does not automatically support every other use.\n\nTranslation requires more than replacing words. Cultural adaptation and testing examine whether questions retain their meaning and whether scores remain comparable across groups."
    },
    {
      "heading": "Collection and scoring",
      "text": "PROs can be collected on paper, by telephone, through web portals, or using mobile devices. Electronic patient-reported outcomes are often abbreviated ePROs. Collection systems may connect with an electronic health record, but electronic delivery alone does not establish measurement quality.\n\nImportant design features include:\n\n- Recall period: whether the patient reports the present moment, the previous day, or a longer period.\n- Assessment schedule: when and how often reports are collected relative to treatment and clinical events.\n- Response options: numerical ratings, verbal categories, or other formats.\n- Administration conditions: the setting, assistance provided, and mode of completion.\n- Scoring rules: how answers are converted into scores and how unanswered items are handled.\n\nThese choices can affect results. A weekly retrospective rating is not necessarily equivalent to an average of daily ratings. Different instruments may also use opposite score directions: a higher score can indicate better functioning on one measure and worse symptoms on another.\n\nComputer-adaptive testing selects questions from a calibrated item bank according to previous answers. It can reduce the number of questions needed for a specified level of measurement precision, but depends on appropriate item calibration and model assumptions."
    },
    {
      "heading": "Interpretation and analysis",
      "text": "A numerical difference is not necessarily important to patients. Statistical significance addresses uncertainty under a statistical model; it does not establish the practical importance of a difference.\n\nMeaningful-change thresholds are often estimated using external anchors, such as a patient’s overall assessment of change, alongside statistical information. Such thresholds depend on the population, instrument, context, and purpose. A threshold for meaningful change within an individual should not automatically be treated as the threshold for an important difference between group averages.\n\nIn a clinical trial, a PRO can serve as a primary, secondary, or exploratory endpoint. Its role, assessment schedule, scoring, and analysis are generally specified in advance. Testing many domains or time points creates multiplicity concerns.\n\nMissing responses are especially important. Patients who become more unwell may be less able to complete questionnaires, so observed scores may not represent all participants. Analyses need to address assumptions about missing data and how events such as treatment discontinuation affect the question being studied. Death is not simply another missing questionnaire response; outcomes involving death require an explicitly defined approach."
    },
    {
      "heading": "Applications",
      "text": "Clinical research and treatment evaluation\n\nPROs help assess whether treatments affect how patients feel and function. They can provide information not captured by survival, imaging, or a biomarker. In cancer research, for example, disease control and patient-reported symptoms describe different aspects of treatment outcomes.\n\nIn randomized controlled trials, PROs can contribute to comparisons of benefits and burdens. Where blinding is not possible, knowledge of treatment allocation may influence reporting and is relevant to interpretation.\n\nRegulatory use requires evidence that an instrument and its interpretation are suitable for the intended population, concept, and claim. A questionnaire’s widespread use does not alone establish that suitability.\n\nClinical care and service evaluation\n\nRepeated PRO collection can document symptom burden and functional change over time. Results may inform clinical discussions and shared decision-making, alongside other evidence.\n\nAggregated PROM data are also used to examine quality of care. Comparisons between providers or services require attention to baseline health, patient characteristics, response rates, and follow-up timing. Differences in scores cannot automatically be attributed to differences in care."
    },
    {
      "heading": "Limitations and implementation challenges",
      "text": "PROs provide direct access to patient experience, but reports can be affected by recall, expectations, wording, language, and the circumstances of completion. Patients may also change their standards or understanding of health over time, a phenomenon known as response shift. This can complicate interpretation of longitudinal scores.\n\nQuestionnaires impose a burden, especially when they are long, repetitive, or poorly matched to patients’ concerns. Literacy, disability, language barriers, and digital access can affect participation and the representativeness of results. Accessible formats and appropriately validated versions are therefore relevant to health equity.\n\nPRO collection also raises data privacy issues because responses may contain sensitive information about symptoms, functioning, or emotional well-being. In clinical services, collection systems need defined arrangements for reviewing results and responding to concerning reports; collecting data is distinct from providing continuous clinical monitoring.\n\nPatient reports and clinical findings may disagree without either being inherently invalid. They can describe different dimensions of illness. A PROM does not independently establish a diagnosis, and a laboratory result does not necessarily describe the burden of a patient’s symptoms."
    }
  ],
  "links": [
    "laboratory-test",
    "pain",
    "fatigue",
    "health-related-quality-of-life",
    "eq-five-d",
    "quality-adjusted-life-year",
    "clinical-trial",
    "clinical-endpoint",
    "biomarker",
    "cancer",
    "randomized-controlled-trial",
    "blinding-clinical-trials",
    "shared-decision-making",
    "quality-of-care",
    "health-equity",
    "data-privacy",
    "diagnosis"
  ],
  "sources": [],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}