Patient Satisfaction

Abstract

Patient satisfaction, which MeSH classifies under treatment adherence and compliance, is the patient's evaluation of care measured against what they expected of it. It is at once an outcome of care, a determinant of whether patients return and adhere, and a contested measure of quality. This article sets out the expectations-disconfirmation account of satisfaction, the structure-process-outcome framework that locates it, and the long argument over whether a global satisfaction score measures care or merely expectation. It distinguishes satisfaction from the specific patient-experience reports that have largely displaced it, surveys what actually drives it, and confronts the finding that higher satisfaction can accompany worse outcomes. A worked example and three interactive demonstrations show identical care yielding opposite satisfaction, the determinants that matter most, and the survey biases that inflate the numbers.

Keywords: patient satisfaction, patient experience, quality of care

For most of medicine's history patient satisfaction was treated as a courtesy, a matter of bedside manner with no bearing on whether care was good. That view has been reversed. Satisfaction is now gathered systematically, tied to reimbursement, and reported as an indicator of health-system quality, on the premise that patients are uniquely positioned to judge aspects of their care that no clinical measure captures. Patient satisfaction names the patient's own evaluation of the care received, formed by comparing what happened against what was expected or hoped for.

The concept is deceptively simple and genuinely contested. A satisfaction score is easy to collect and hard to interpret: it reflects the care, but also the patient's prior expectations, their disposition, and the way the question was asked. The sections below set out what satisfaction is, how it is measured, what drives it, how it relates to clinical outcomes, and why a measure everyone now collects remains one the field cannot fully agree on how to read.

What Patient Satisfaction Is

Patient satisfaction is an evaluative judgment: the patient weighs the care received against a standard and reports how well it measured up. Sonja Linder-Pelz gave the idea its first explicit theory, arguing that satisfaction is an attitude, a set of evaluations shaped by the patient's expectations and the value they place on each aspect of care, rather than a direct readout of care quality (Linder-Pelz, 1982). On this account a satisfaction report cannot be understood without knowing the expectation it was measured against, because the same objective care produces different satisfaction in patients who expected different things.

Gregory Pascoe sharpened the mechanism into a comparison process: patients judge care by contrasting their experience with a cognitive standard, and satisfaction is the result of that disconfirmation, positive when experience exceeds the standard and negative when it falls short (Pascoe, 1983). John Ware and colleagues then established that satisfaction is not one thing but several, identifying distinct dimensions, interpersonal manner, technical quality, accessibility, finances, and more, that patients evaluate separately, so that a person can be delighted with their doctor's warmth and dismayed by the wait (Ware et al., 1983). This multidimensional, expectation-relative structure is what makes satisfaction both informative and difficult: it carries real signal about care, entangled with signal about the patient.

Types of Patient Satisfaction

In the Medical Subject Headings vocabulary, patient satisfaction sits among the headings used to index how patients relate to and evaluate their care, and it has one narrower descriptor directly beneath it. MeSH is an indexing classification, not a theory of the construct, so the placement below reflects how the literature is catalogued rather than a claim that satisfaction decomposes cleanly into this single child; the dimensions of satisfaction described above cut across it.

The single child understates how much structure satisfaction has. The more consequential subdivision is the one the measurement literature draws, between global satisfaction, a single overall verdict, and satisfaction with specific dimensions of care, each of which can move independently (Ware et al., 1983). It is that dimensional structure, not the MeSH tree, that tells a hospital which part of its care a low score indicts.

Measuring Satisfaction

If satisfaction is multidimensional and expectation-relative, measuring it well is a serious methodological problem, and the field spent two decades on it. Ray Fitzpatrick's generation of survey work gave way to systematic appraisal: Josephine Sitzia and Neil Wood reviewed the concept and found the literature plagued by weak definitions and instruments built without psychometric discipline, so that many reported satisfaction rates were uninterpretable (Sitzia & Wood, 1997). Rosamund Crow and colleagues, in a systematic review for the UK Health Technology Assessment programme, documented the same problem at scale and set out what a defensible measure requires, a clear construct, tested reliability and validity, and attention to the expectations that confound raw scores (Crow et al., 2002).

The most important methodological turn was away from satisfaction itself. Crispin Jenkinson and colleagues showed that asking patients to rate how satisfied they were produced uniformly high, poorly discriminating scores, whereas asking what specifically happened, whether a doctor explained things clearly, whether pain was controlled, whether the patient was treated with dignity, produced reports that varied, discriminated between providers, and pointed at fixable failures (Jenkinson et al., 2002). This is the distinction between satisfaction and patient experience, and it reframed measurement around concrete, reportable events rather than global affect.

The care is fixed — communication 70, waiting 40, thoroughness 80. Only the patient’s expectations and weights change. Each aspect scores 50 when the care exactly meets expectation, higher when it beats it, lower when it falls short; overall satisfaction is the weighted sum.

The two presets receive precisely the same care yet report 29 and 68 — more than two-fold apart. A clinic reading these as a verdict on its care would conclude it failed one patient and delighted the other, when it did the same thing for both. The expectation, not the care, moved the score.

What Drives Satisfaction

If satisfaction is an evaluation, the practical question is what it responds to. Judith Hall and Michael Dornan's meta-analysis of the satisfaction literature answered it quantitatively: across dozens of studies, the aspects of care that most strongly predicted satisfaction were interpersonal, how much information the clinician gave and the warmth and respect of their manner, outweighing the technical and structural features that systems more readily measure (Hall & Dornan, 1988). Satisfaction, in other words, is earned largely in the conduct of the encounter.

A later systematic review by Enkhjargal Batbaatar and colleagues confirmed and broadened the picture, organizing the determinants into patient factors (expectations, health status, sociodemographics) and care factors (interpersonal manner, communication, technical competence, accessibility, and the physical environment), with communication and the clinician-patient relationship recurring as the dominant modifiable drivers (Batbaatar et al., 2017). The determinants split into those a health system can change and those it cannot, which is why interpretation matters: a low score driven by high expectations calls for a different response than one driven by poor communication.

Five modifiable drivers of satisfaction, weighted by how strongly the evidence says each predicts it. Set each driver’s performance; the bar shows its contribution and the note shows how much a ten-point gain there adds to the overall score. Interpersonal drivers carry the most weight.

A fixed improvement in communication or manner raises satisfaction several times more than the same improvement in amenities. Meta-analysis finds satisfaction is earned largely in the conduct of the encounter, which is why how care is delivered outweighs much of what is delivered.

Satisfaction, Experience, and Outcomes

Whether satisfaction tracks the quality of care is the question the whole enterprise rests on, and the evidence is genuinely mixed. On one side, Chris Doyle, Laura Lennox, and Derek Bell's systematic review found consistent positive associations between patient experience and both clinical safety and effectiveness across a wide range of conditions, supporting the case that experience is a legitimate pillar of quality rather than a soft adjunct to it (Doyle et al., 2013). Matthew Manary and colleagues argued in the same spirit that well-measured patient experience carries real information about care processes and should be taken seriously as an outcome (Manary et al., 2013).

On the other side sits a genuinely uncomfortable finding. Joshua Fenton and colleagues, following a national sample, found that the most satisfied patients had greater healthcare expenditure, more inpatient admissions, and, strikingly, higher mortality, suggesting that satisfaction can be bought with acquiescence to patient requests that are not in the patient's clinical interest (Fenton et al., 2012). Rebecca Anhang Price and colleagues reconciled the tension by arguing that patient experience measures capture a distinct and valid domain of quality, neither redundant with clinical measures nor a substitute for them, and that their value depends entirely on measuring experience rather than raw satisfaction (Anhang Price et al., 2014). The lesson is consistent: satisfaction is a signal to interpret alongside clinical quality, never a proxy to maximize in its place.

Worked Example: Identical Care, Opposite Satisfaction

Consider two patients who receive exactly the same care at the same clinic. Three aspects of that care are scored for perceived performance on a 0-to-100 scale: communication (70), waiting time (40), and clinical thoroughness (80). These numbers are fixed; the care does not vary between the two patients.

What varies is expectation. Following the disconfirmation account, each attribute's satisfaction contribution is `50 + (performance − expectation)`, clamped to the 0-to-100 range, so meeting expectation yields a neutral 50, exceeding it rises above, and falling short drops below. Each patient also weights the three attributes, and overall satisfaction is the weighted sum.

- Patient 1 arrives with high expectations: communication 90, waiting 80, thoroughness 90; weights 0.5, 0.2, 0.3. - Patient 2 arrives with modest expectations: communication 50, waiting 30, thoroughness 60; weights 0.5, 0.2, 0.3.

Patient 1's contributions are communication `50 + (70 − 90) = 30`, waiting `50 + (40 − 80) = 10`, thoroughness `50 + (80 − 90) = 40`, giving `0.5×30 + 0.2×10 + 0.3×40 = 15 + 2 + 12 = 29`. Patient 2's are communication `50 + (70 − 50) = 70`, waiting `50 + (40 − 30) = 60`, thoroughness `50 + (80 − 60) = 70`, giving `0.5×70 + 0.2×60 + 0.3×70 = 35 + 12 + 21 = 68`. The care was identical; the satisfaction scores, 29 and 68, are more than two-fold apart, driven entirely by the expectations each patient brought. A clinic reading these scores as a verdict on its care would conclude it had failed Patient 1 and delighted Patient 2, when it had done precisely the same thing for both (Linder-Pelz, 1982; Fenton et al., 2012).

Discussion

The study of patient satisfaction has had to hold two truths together. The first is that patients know things about their care that no instrument and no clinician can supply, whether they were treated with dignity, whether their questions were answered, whether they understood their options, and that gathering these reports systematically has improved care. The second is that a satisfaction score is a noisy, expectation-laden signal that rewards the wrong things if treated as a target (Fenton et al., 2012). The field's maturation has been the move from collecting satisfaction to measuring experience, from asking patients how they felt to asking what happened (Jenkinson et al., 2002; Anhang Price et al., 2014).

This connects satisfaction to the broader architecture of quality. Avedis Donabedian's framework, which assesses care along the triad of structure, process, and outcome, located satisfaction as both an outcome of care and a contributor to it, since a satisfied patient is more likely to return, to disclose, and to adhere (Donabedian, 1988). That link to treatment adherence is why MeSH files satisfaction where it does, and why the measure matters beyond the survey: care a patient is satisfied with is care they are more likely to continue.

Cognitive Implications

Patient satisfaction is, at bottom, a judgment under comparison, and it inherits the properties of such judgments. The disconfirmation model that Linder-Pelz and Pascoe brought into healthcare was formalized a few years earlier by Richard Oliver as a general cognitive account of satisfaction: a person compares a product or service against a prior expectation, and it is the disconfirmation of that expectation, not the experience in isolation, that determines satisfaction (Oliver, 1980). This is a special case of the expectation-contrast effects studied throughout decision making: the same stimulus is evaluated more favorably against a low reference point than a high one, so the reference point does as much work as the stimulus (Linder-Pelz, 1982; Pascoe, 1983). A satisfaction rating is therefore not a measurement of care in the way a blood pressure is a measurement of pressure; it is a constructed evaluation in which the standard of comparison is part of what is measured.

This has a sharp methodological consequence that Benjamin Williams pressed: if satisfaction is constructed from expectations, and patients often hold low expectations of what they are entitled to, then high satisfaction can coexist with poor care, because patients do not judge against a standard they never held (Williams, 1994). The cognitive point, that evaluation is relative to an internal reference, is also the critical point, and it is why the determinants of a score include the patient's mind as well as the clinic's conduct. Good measurement works around this by asking about concrete experience, which has less room for the reference point to float, rather than about global feeling.

Current Directions

The active frontier is the measurement problem made rigorous. A systematic review by Claudia Bull and colleagues examined the validity and reliability of patient-reported experience measures and found the psychometric quality of many instruments still under-tested, so that the shift from satisfaction to experience has not yet been matched by uniformly sound tools, which remains the field's central methodological task (Bull et al., 2019).

A second current concerns bias. Felipe Dunsch and colleagues showed experimentally that the framing of satisfaction questions, in particular positive-agreement wording, inflates reported satisfaction and threatens the validity of the measure, a courtesy-and-acquiescence bias that specific-experience questions only partly escape (Dunsch et al., 2018). Running alongside this is a conceptual argument advanced by Elizabeth Larson and colleagues for the World Health Organization: that patients should be treated as expert reporters of their own care, and that measurement should be designed to capture that expertise cleanly rather than to produce a flattering number (Larson et al., 2019). The open questions are practical, how to word and weight instruments so they discriminate real differences in care, how to separate the patient's expectation from the clinic's performance, and how to use the resulting scores to improve care rather than merely to rank it.

Set the true quality of care and how strong the courtesy-and-acquiescence bias is. A global “how satisfied were you?” item is pulled toward the ceiling by that bias; a specific-experience item that asks what actually happened resists most of it. The gap is the inflation a single satisfaction number hides.

As the bias rises, the global score climbs toward 100 almost regardless of the care, which is why satisfaction scores cluster high and discriminate poorly between providers. Asking about concrete events leaves far less room for the courtesy effect, which is the measurement case for experience over satisfaction.

Common Misconceptions

A high satisfaction score means good care.

Not reliably. Satisfaction is judged against expectations, so patients with low expectations can report high satisfaction for poor care, and the most satisfied patients in one national study had higher mortality (Williams, 1994; Fenton et al., 2012).

Satisfaction and patient experience are the same thing.

They are distinct. Satisfaction is a global evaluative feeling; experience is a report of what specifically happened. Experience questions discriminate between providers and point at fixable failures where satisfaction scores cluster uninformatively high (Jenkinson et al., 2002).

Satisfaction is mostly about clinical outcome.

It is mostly about the encounter. Meta-analysis finds that how a clinician communicates and the warmth of their manner predict satisfaction more strongly than technical or structural features of care (Hall & Dornan, 1988).

A single overall satisfaction number is enough.

No. Satisfaction is multidimensional, and a global score hides which dimension, communication, access, interpersonal manner, or technical care, actually failed (Ware et al., 1983).

Glossary

Acquiescence bias.

The tendency of respondents to agree with positively worded survey items, inflating reported satisfaction and threatening the validity of the measure.

Courtesy bias.

The tendency of patients to give favorable evaluations out of politeness or gratitude toward their caregivers, independent of the care's actual quality.

Disconfirmation.

The gap between experienced care and a prior expectation; a positive gap raises satisfaction and a negative gap lowers it, in the comparison model of satisfaction.

Expectation.

The standard a patient holds for what care should be, against which the care received is evaluated; a central confound in interpreting satisfaction scores.

Global satisfaction.

A single overall evaluation of care, which tends to cluster high and discriminate poorly between providers.

Interpersonal manner.

The warmth, respect, and attentiveness with which a clinician treats a patient; the strongest modifiable driver of satisfaction in meta-analysis.

Multidimensionality.

The property that satisfaction comprises separable components, interpersonal manner, communication, access, and technical quality, that can move independently of one another.

Outcome.

In Donabedian's triad, the effects of care on a patient's health and welfare, among which satisfaction is itself counted.

Patient experience.

A report of what specifically happened during care, such as whether a doctor explained things clearly, as distinct from how satisfied the patient felt.

Patient preference.

The relative desirability a patient assigns to the options in a health decision; the narrower MeSH descriptor beneath patient satisfaction.

Patient satisfaction.

The patient's evaluation of care, formed by comparing the care received against prior expectations, and treated as both an outcome and a quality indicator.

Patient-reported experience measure (PREM).

A standardized instrument that asks patients about concrete events in their care, designed to be more actionable and less expectation-dependent than a satisfaction rating.

Process.

In Donabedian's triad, what is actually done in giving and receiving care; the component to which satisfaction most directly speaks.

Structure-process-outcome.

Donabedian's framework for assessing healthcare quality along the resources of care, the activities of care, and their results, within which satisfaction is both an outcome and a contributor.

Structure.

In Donabedian's triad, the settings, equipment, staffing, and resources within which care is delivered.

Key Researchers

Paul D. Cleary

(Yale School of Public Health). Health-services researcher central to the development of standardized patient-experience measurement, including the surveys that turned satisfaction from a local questionnaire into a national instrument of health-system accountability, and who reframed patient reports as measures of care processes patients are uniquely positioned to judge. ORCID · Yale

Angela Coulter

(University of Oxford). Health-services researcher who made patient-centred care and the systematic measurement of patient experience a mainstream policy concern, arguing that eliciting and acting on what patients report is a clinical obligation rather than a courtesy. ORCID

Avedis Donabedian

(University of Michigan, 1919–2000). The foundational figure in healthcare quality assessment, whose structure-process-outcome triad gave the field its organizing framework and located patient satisfaction as both an outcome of care and a contributor to it. Wikipedia · Wikidata

Judith A. Hall

(Northeastern University). Social psychologist of nonverbal communication and the physician-patient interaction, whose 1988 meta-analysis of satisfaction with medical care first quantified how provider behavior, information-giving and interpersonal manner, drives what patients report. Google Scholar

Crispin Jenkinson

(University of Oxford). Health-services researcher and authority on patient-reported outcome and experience measures, whose questionnaire studies of specific aspects of care distinguished the dimensions of experience that global satisfaction ratings obscure. Oxford NDPH

John E. Ware Jr.

(UMass Chan Medical School). Health-outcomes measurement scientist who, before creating the SF-36, co-authored the early work defining and measuring patient satisfaction, establishing the multidimensional, psychometrically disciplined approach that separated satisfaction measurement from ad hoc survey practice. UMass Chan

Frequently Asked Questions

What is patient satisfaction?

It is the patient's own evaluation of the care they received, formed by comparing that care against what they expected or hoped for. It is treated as both an outcome of care and an indicator of quality, and MeSH classifies it beneath treatment adherence and compliance (Linder-Pelz, 1982).

How is patient satisfaction measured?

Through surveys, but good measurement has shifted from asking how satisfied a patient felt to asking what specifically happened, because specific-experience questions discriminate between providers and point at fixable failures where global satisfaction scores cluster uninformatively high (Jenkinson et al., 2002; Crow et al., 2002).

What is the difference between satisfaction and patient experience?

Satisfaction is a global evaluative feeling; experience is a report of concrete events, such as whether a doctor explained things clearly or whether pain was controlled. Experience measures are more actionable and less dependent on the patient's prior expectations (Anhang Price et al., 2014).

What drives patient satisfaction most?

Interpersonal conduct. Meta-analysis finds that how much information a clinician gives and the warmth and respect of their manner predict satisfaction more strongly than amenities or even technical features of care (Hall & Dornan, 1988; Batbaatar et al., 2017).

Does higher satisfaction mean better care?

Not necessarily. One large national study found the most satisfied patients had higher expenditure, more admissions, and higher mortality, suggesting satisfaction can be bought with acquiescence to requests that are not clinically beneficial. It is a signal to interpret, not a target to maximize (Fenton et al., 2012).

Why can the same care produce different satisfaction scores?

Because satisfaction is judged against expectation. A patient with high expectations and one with modest expectations can receive identical care and report sharply different satisfaction, since each compares the care to a different internal standard (Linder-Pelz, 1982; Williams, 1994).

Is a single overall satisfaction score enough?

No. Satisfaction is multidimensional, so a global number hides which dimension, communication, access, interpersonal manner, or technical care, actually succeeded or failed. Dimension-level measurement is needed to act on a score (Ware et al., 1983).

Does patient experience relate to clinical quality?

Yes. A systematic review found consistent positive associations between patient experience and both clinical safety and effectiveness across many conditions, supporting experience as a legitimate component of quality rather than a soft adjunct (Doyle et al., 2013).

References