The Journal of Emergency Medicine, Vol. 26, No. 1, pp.

13–26, 2004
Copyright © 2004 Elsevier Inc.
Printed in the USA. All rights reserved
0736-4679/04 $–see front matter

doi:10.1016/j.jemermed.2003.04.003

Original
Contributions

PATIENT SATISFACTION IN THE EMERGENCY DEPARTMENT: A REVIEW OF

THE LITERATURE AND IMPLICATIONS FOR PRACTICE
Edwin D. Boudreaux, PhD* and Erin L. O’Hea, PhD†

*Department of Emergency Medicine, University of Medicine and Dentistry of New Jersey-Robert Wood Johnson Medical School and
Cooper Hospital, Camden, New Jersey, and †Department of Psychiatry, University of Medicine and Dentistry of New Jersey-Robert
Wood Johnson Medical School, Piscataway, New Jersey
Reprint Address: Edwin D. Boudreaux, PHD, Emergency Medicine Residency Program, Cooper Hospital, One Cooper Plaza, Camden,
NJ 08103

e Abstract—This article reviews the empirical literature perceived waiting times. To advance this area of research,
on patient satisfaction in the Emergency Department (ED). investigators must use: 1) larger, more representative sam-
It explores the implications for clinical practice, discusses ples; 2) reliable and valid assessment instruments; 3) theo-
limitations and weaknesses of the literature, and provides ry-driven hypothesis testing; and 4) randomized, controlled
direction for future research. Articles resulting from a trials. © 2004 Elsevier Inc.
comprehensive electronic search were obtained, their ref-
erences examined, and all other relevant articles not al- e Keywords—patient satisfaction; emergency medicine;
ready discovered via the electronic search were acquired emergency department; interpersonal skills; acuity; length
and reviewed. Articles were included if: 1) the stated goal of of stay; technical skills; waiting time
the study was to investigate satisfaction with at least one
aspect of ED care, 2) the study was conducted in the United
States, 3) it provided enough information on the study
methods, design, and statistical analyses to conduct a crit- INTRODUCTION
ical review, and 4) it used quantitative methods. Fifty stud-
ies met the above criteria. Based on the multivariate pre- Patient satisfaction deserves attention not only because it
dictive studies, the most robust predictor of global is an intrinsically worthy goal, but also because it is a
satisfaction is the quality of interpersonal interactions with potentially significant mediator for a range of important
the ED provider. Perceived waiting times are more closely outcomes. Satisfied patients may be more compliant with
associated with satisfaction than actual waiting times. Sev- their medical regimens, suggesting that satisfaction may
eral methods for improving satisfaction have shown prom- be an important component in promoting health and
ise, but none has garnered sufficient support to recommend well-being (1–3). Satisfaction may also directly impact
unequivocally. Promising interventions include: providing
the financial viability of an institution by affecting con-
information on how the ED functions through visual media,
improving ED processes through performance improve-
sumer choice, with dissatisfied patients choosing to
ment methodologies, and improving the interpersonal skills “shop” for healthcare services elsewhere. These factors,
of providers. Interventions designed to reduce actual wait- combined with a growing emphasis on provider account-
ing times have not been sufficiently studied, but results ability, competition for a restricted number of healthcare
from several well-designed studies suggest that such a strat- dollars, and a desire to reduce professional liability
egy is unlikely to have as great an impact as those targeting claims, have led to a proliferation of studies and com-

RECEIVED: 18 September 2001; FINAL SUBMISSION RECEIVED: 25 February 2003;

ACCEPTED: 8 April 2003
13
14 E. D. Boudreaux and E. L. O’Hea

Table 1. Characteristics of Observational Multivariate Studies

Study ED type Volume Sample size Assessment method and time

(% eligible pts) post-ED visit

Baker et al. (29) Public; urban; 500⫹ bed NA 467 (61.0%) Tel., 7–11 days
hospital
Boudreaux et al. (30) Academic; urban 85,000/yr 437 (38.5%) Tel., 7–10 days
Bursch et al. (31) Non-academic; HMO; 331-bed NA 258 (59.0%) Tel., 7 days
hospital
Campanella et al. (32) Military; 250-bed hospital NA 178 (33.0%) Mail, 4–7 days
Carey and Seibert (33) 42% academic; 30% religious; NA 17,271 (33%) Mail, 7 days
53% 300⫹ bed hospitals
Hall and Press (34) Mixed 7915/yr to 81,834/yr; 3085 (⬃25.0%)Mail, 3–4 days
x ⫽ 33,505/yr
Mack et al. (35) Mixed NA 493 (38.0%) Tel., random digit dialing, ED visit
in past 12 mo
Raper (36) Academic; urban NA 200 (50.0%) On-site survey; tel., 1–2 days
Raper et al. (37) 1 ⫽ Academic; urban; Level I 1 ⫽ 40,000/yr 378 (??%) Tel., 1–2 days
trauma center 2 ⫽ Private; 2 ⫽ 56,000/yr
non-profit 3 ⫽ Private; 3 ⫽ 43,000/yr
religious
Rhee and Bird (38) Academic; urban; Level I 23,000/yr 618 (46%) Tel., 6 days
trauma center
Sun et al. (39) Academic; urban; non-profit NA 2,333 (67.5%) On-site survey; tel., 7–12 days
Sun et al. (40) Academic; urban; non-profit NA 2,373 (22.9%) Mail, 2–10 days; FU mail 14–24
days
Yarnold et al. (41), A Academic; urban; Level I 48,000/yr 2198 (17%) Mail, 7 days
trauma center
Yarnold et al. (41), B Community; suburban 40,000/yr 1631 (45%) Tel., 12–28 days

NA ⫽ not available; Tel. ⫽ telephone; FU ⫽ follow up; yr ⫽ year.

mentaries on patient satisfaction over the last several ined manually, and all other relevant articles not discov-
decades. ered via the electronic search were acquired.
Although the specialty of Emergency Medicine is
relatively new, it has not been overlooked in the patient
satisfaction movement. However, like the larger patient Criteria for Review
satisfaction literature, many of the existing Emergency
Department (ED) patient satisfaction studies have seri- Studies were retained for the current review if they met
ous methodological problems, which has led to inconsis- four criteria. First, we only included studies that were
tent and, at times, contradictory conclusions. In this primarily focused on patient satisfaction. Second, we
article, we review the most methodologically sound stud- chose to focus only on studies conducted in the United
ies in this field to answer two questions: 1) what are the States because of the difficulties inherent in comparing
strongest predictors of ED patient satisfaction? and 2) results derived from countries with different healthcare
what changes can be made to improve ED patient satis- systems (i.e., free market versus nationalized healthcare).
faction? Based on our review, practical recommenda- Third, we excluded studies that failed to provide enough
tions for clinical practice and suggestions for future details to critically evaluate the study methods, research
investigations are made. design, statistical analyses, and results. Finally, each
study must have used quantitative research methods as
defined by: 1) sampling of a general ED population or a
clearly delineated sub-population (e.g., patients present-
METHODS
ing to the ED with chest pain), 2) use of a standardized
Review of the Literature measure of patient satisfaction, 3) description of the
statistical methods used to analyze the data, and 4) de-
We searched three electronic databases (MEDLINE, CI- scription of the results of the statistical analyses, includ-
NAHL, and PSYCH-INFO) using the keywords “patient ing effects sizes and p-values where appropriate.
satisfaction,” “customer satisfaction,” “emergency de- The articles that met the above criteria were then
partment,” and “emergency medicine.” No date or lan- classified into three categories based on research design:
guage restrictions were used. Articles resulting from this observational, univariate; observational, multivariate;
electronic search were retrieved, their references exam- and interventional. Finally, the manner in which both
Emergency Department Patient Satisfaction 15

Table 2. Methodological Review of Observational Multivariate Studies*

Study Internal External Study measure Predictor Statistics Average

representativeness representativeness inclusion

1 1 3 1 3 2.0
Baker et al. (29) Non-urgent only Single site 40% Reported No assessment of Linear and logistic
7:00–23:00 only Hispanic psychometrics tech. skills, regressions
No comparison Previously perceived
b/t resp. and validated Q. waiting times,
non-resp. actual waiting
times, or acuity
1 1 2 3 1 1.6
Boudreaux (30) Low response rate Single site State Reported Broad range of Logistic regression
Non-urgent and supported psychometrics predictors No control for
female were municipal ED Created for multicolinearity
over- study
represented
1 1 1 2 1 1.2
Bursch et al. (31) No comparison b/t High acuity HMO No report of No assessment of Linear regression
resp. and non- participants psychometrics actual waiting No control for
resp. times or tech. multicolinearity
skills
1 1 3 2 3 2.2
Campanella et al. No admitted pts. Single site Military Widely used Q. No assessment of Linear regression
(32) No comparison hospital with race or acuity Accounted for
b/t resp. and established multicolinearity
non-resp. psychometrics
1 3 3 1 3 2.2
Carey and Seibert No comparison b/t 235 EDs; 43 US Widely used Q. No assessment of Logistic regression
(33) resp. and non- states and all 8 with age, sex, race, Accounted for
resp. census regions established actual waiting multicolinearity
psychometrics times, or acuity
1 3 3 2 3 2.6
Hall and Press (34) Low response rate 23 EDs; 13 US Widely used Q. No assessment of Linear regression
states with race or acuity Accounted for
established multi-colinearity
psychometrics
1 2 1 1 2 1.4
Mack et al. (35) Privately insured 5 cities No report of No assessment of Linear regression
pts. only psychometrics age, sex, race, Provided
created for perceived intercorrelations
study waiting times, but did not
or actual control for
waiting times multicolinearity
1 1 2 2 2 1.4
Raper (36) Restricted CC No Single site Q. with No assessment of Linear regression
comparison b/t established perceived Provided
resp. and non- psychometrics, waiting times or intercorrelations
resp. but poorly actual waiting but did not
described in times control for
article multicolinearity
1 2 2 1 2 1.6
Raper et al. (37) Restricted CC No 3 EDs in 2 cities Q. with No assessment of Linear regression
comparison b/t established tech. skills, Provided
resp. and non- psychometrics, perceived intercorrelations
resp. but poorly waiting times, but did not
described in or actual control for
article waiting times multicolinearity
3 1 2 1 1 1.6
Rhee and Bird (38) Random selection Single site Reported No assessment of Linear regression
Compared resp. psychometrics sex, actual No control for
and non-resp. Created for waiting times, multi-colinearity
study or acuity
(Continued)

predictors and criterion variables were measured or de- from psychometrically sound, widely used question-
fined in the existing research varied considerably by naires, whereas others used questionnaires specifically
study. Some studies derived their predictor variables created for the study. Similarly, some studies have used
16 E. D. Boudreaux and E. L. O’Hea

Table 2. continued

Study Internal External Study measure Predictor Statistics Average

representativeness representativeness inclusion

2 2 3 2 3 2.4
Sun et al. (39) Selected CCs 5 EDs in Boston Reported No assessment of Logistic regression
Comparison b/t psychometrics tech. skills Accounted for
resp. and non- Used in multicolinearity
resp. previous by use of
studies theoretically
independent
var.’s
2 2 3 1 3 2.2
Sun et al. (40) Selected CCs 4 EDs in Boston Reported No assessment of Logistic regression
Comparison b/t psychometrics sex, tech. skills, Accounted for
resp. and non- Used in acuity, or actual multicolinearity
resp. previous waiting times by use of
studies theoretically
independent
var’s
1 1 3 1 3 1.8
Yarnold et al. (41), A Low response rate Single site Widely used Q No assessment of Classification tree
No comparison with age, sex, race, analysis
b/t resp. and established acuity, or actual Accounted for
non-resp. psychometrics waiting times multicolinearity
1 1 1 1 3 1.4
Yarnold et al. (41), B Low response rate Single site No reported No assessment of Classification tree
No comparison psychometrics age, sex, race, analysis
b/t resp. and Not widely acuity, or actual Accounted for
non-resp. used Q. waiting times multicolinearity

Q ⫽ questionnaire; tech ⫽ technical; resp. ⫽ responders; b/t ⫽ between; pts. ⫽ patients; CCs ⫽ chief complaints.
* The following rating system was applied to rate the methodology of each study: Internal Representativeness (how well does the sample
represents the general ED population from which it was drawn?): 3 ⫽ excellent representativeness; not focused on specific subsample;
compares those enrolled to those not enrolled; 2 ⫽ good representativeness; may focus on large subsample (e.g., discharged patients);
compares those enrolled to those not enrolled; 1 ⫽ poor representativeness; focus on small subsample (e.g., only those with certain
presenting complaints); no comparison of those enrolled to those not enrolled. Generalizability to Other EDs (how well does the sample
represent United States EDs as a whole?): 3 ⫽ excellent, multi-site, national sample; 2 ⫽ good, multi-site, regional sample; 3 ⫽ poor,
single-site sample. Characteristics of Study Measure (how psychometrically sound is the instrument?). 3 ⫽ excellent, strong reliability
and validity data reported in article or questionnaire widely used and psychometrics reported elsewhere; 2 ⫽ good, acceptable reliability
and validity data reported, but questionnaire was specifically constructed for the study (i.e., not previously or widely used); 1 ⫽ poor,
unacceptable (or no) reliability and validity data reported. Predictor Inclusion (how well does the study compare across a range of
relevant variable domains?): 3 ⫽ excellent, includes (at a minimum) items assessing interpersonal interactions with provider, perceived
provider technical skill, perceive waiting times, actual waiting times, age, gender, race, and acuity; 2 ⫽ good, includes most but not all
of those listed in “excellent” category; 1 ⫽ poor, greatly restricted inclusion. Statistical Analyses (are the statistics appropriate and do
they account for multicolinearity?): 3 ⫽ excellent, appropriate parametric or nonparametric statistics used; explicitly addressed issue of
multicolinearity through use of factor scores or independent predictors; 2 ⫽ good, appropriate statistics used and did not explicitly
address issue of multicolinearity but did report intercorrelations among predictors; 1 ⫽ poor, used questionably appropriate statistics
and did not address multicolinearity.

ratings of overall satisfaction as the criterion variable, criterion variable was used (overall satisfaction, likeli-
whereas others have used ratings of the likelihood to hood to recommend, intention to return).
recommend the ED to others or intention to return for
future emergency care. The use of different predictors
and criterion variables makes comparisons across studies Methodological Evaluation of Studies
difficult. To help guide our interpretations, we classified
the statistically significant predictors of each study into The articles meeting the above criteria were evaluated
the following predictor domains: interpersonal interac- independently by the authors using a standardized rating
tion with providers, perceived technical skills of provid- procedure created for the purposes of this paper. A
ers, perceived waiting times, actual waiting times, pa- three-point score (1 ⫽ poor, 2 ⫽ good, 3 ⫽ excellent)
tient characteristics, visit characteristics, and facility was assigned to each of the following areas: internal
characteristics. The legend of Table 3 contains examples representation (how well does the sample represent the
of the individual predictor variables included under each general ED population from which it was drawn?), ex-
predictor domain. Finally, we specified in Table 3 which ternal representation (how well does the sample repre-
Emergency Department Patient Satisfaction 17

Table 3. Conclusions of Non-interventional, Multivariate Studies

Study Criterion variable Conclusions Variable domain* with

largest effect size

Baker et al. (29) Interpersonal aspects Spanish-speaking patients who communicate with Interpersonal interaction
of care use of an interpreter were less satisfied with with providers
interpersonal aspects of care than those who (ethnicity/interpreter
could communicate adequately without one; use)
those who did not communicate with an
interpreter but who believed they should have
had one present were the least satisfied.
Boudreaux et al. (30) Overall satisfaction The degree to which the patient was treated as a Overall satisfaction:
Likelihood to person, feelings of safety and security, and Interpersonal
recommend being provided understandable DC instructions interaction with
were the most important predictors of overall providers
satisfaction and likelihood to recommend. Likelihood to
Perceived technical skills of the RN and wait recommend:
time prior to MD assessment were also Interpersonal
important for overall satisfaction, while interaction with
insurance status, age, and expressive quality of providers
RN staff were important for likelihood to
recommend.
Bursch et al. (31) Overall satisfaction Satisfaction with wait for care, expressive quality, Perceived wait time
organization skills of staff, and information
delivery were most important predictors of
overall satisfaction. Demographics and other
visit characteristics were unimportant.
Campanella et al. (32) Likelihood to Staff interpersonal and communication skills, Interpersonal interaction
recommend satisfaction with wait, and perceptions of the with providers
care being worth the money are important
predictors of likelihood to return.
Carey and Seibert (33) Overall satisfaction RN and MD care are stronger predictors of Interpersonal interaction
satisfaction than facility, testing services, with providers
perceived waiting times, medical outcome, and
registration.
Hall and Press (34) Likelihood to RN and MD care are stronger predictors of Interpersonal interaction
recommend likelihood to recommend than convenience/wait with providers
times, ED census, length of stay, age, and sex.
Mack et al. (35) Likelihood to Satisfaction with medical care and staff Likelihood to
recommend interactions are more important than recommend: technical
Intention to return satisfaction with the facility in determining skills of the providers
likelihood to recommend and intention to return. Intention to return:
Patients brought in by ambulance have higher visit characteristics
intention to return but equivalent likelihood to (arrival by ambulance)
recommend to others.
Raper (36) Satisfaction with RN Interpersonal skills of RN (i.e., psychological Satisfaction with RN
care safety and information giving) predicted Care: interpersonal
satisfaction with RN care, while discharge interaction with
teaching and perceived technical skills did not. provider
Satisfaction with RN care predicted intention to
return for future emergencies.
Raper et al. (37) Satisfaction with Race of triage RN, education level of triage RN, Satisfaction with triage
triage care and satisfaction with triage RN predicted care: interpersonal
Intention to return satisfaction with triage care. Satisfaction with interaction with
triage RN, type of hospital (academic and provider
religious), older age, and level of RN education Intention to return:
predicted greater intention to return in future interpersonal
emergencies. interaction with
provider
Rhee and Bird (38) Overall satisfaction Patient perceptions of provider technical skills Technical skills of
were stronger predictors of overall satisfaction providers
than perceived timeliness and interpersonal
interactions with providers.
Sun et al. (39) Overall satisfaction Actual wait to see MD and total length of stay are Overall satisfaction and
Likelihood to not important predictors of satisfaction; likelihood to
recommend responsiveness of staff and communication of recommend:
information (e.g., causes of problem, reasons Interpersonal
behind wait, DC instructions) were the most interaction with
important predictors of satisfaction. Black race providers
and younger age may be associated with
greater dissatisfaction.
(Continued)
18 E. D. Boudreaux and E. L. O’Hea

Table 3. continued

Study Criterion variable Conclusions Variable domain* with

largest effect size

Sun et al. (40) Overall satisfaction Validated conclusions from Sun et al., 2000. Interpersonal interaction
Communication of information (e.g., causes of with providers
problem, reasons behind wait, DC instructions)
most important predictors of satisfaction.
Yarnold (41), A and B Overall satisfaction Expressive quality of MD and RN are strongest Interpersonal interaction
predictors of satisfaction. Perceived waiting with providers
times also can be important.

RN ⫽ nurse; MD ⫽ physician; DC ⫽ discharge.

* Predictor domains: Interpersonal interaction with providers (e.g., expressive quality, information delivery, bedside manner, respon-
siveness, availability); technical skills of providers; perceived wait times; actual wait times; patient characteristics (e.g., age, sex, race,
ethnicity, insurance status); visit characteristics (e.g., acuity, diagnosis, disposition); and facility characteristics (e.g., ED daily census,
annual volume, hospital type).

sent the EDs in the United States as a whole?), study Significant Predictors of ED Patient Satisfaction
measure (how psychometrically sound is the instru-
ment?), study design (multivariate studies: how well To answer the first research question, we relied heavily
does the study compare across a range of relevant pre- upon multivariate studies with ‘good’ or ‘excellent’ pre-
dictor domains? interventional studies: how well does dictor inclusion (studies with a ‘2’ or ‘3’ for Predictor
the study control for confounding factors?), and statisti- Inclusion in Table 2). This was done because the univar-
cal analyses (are the statistics appropriate for the variable iate studies, and many of the less-impressive multivariate
distribution and the research question?) Differences in studies, only investigated the relation between one do-
ratings between the two raters were reconciled via con- main, such as waiting times, and ignored other poten-
ference. More details on this rating system, including the tially important domains, such as interactions with pro-
specific criteria for each rating area, are listed in Tables vider. Hence, this compromised many researchers’
2 and 5. ability to make conclusions regarding which predictors
were most strongly linked to patient satisfaction, because
these studies did not include a broad range of potentially
RESULTS important predictors.

Of the 107 studies identified by the electronic search and

the manual bibliography review, 57 articles were elimi- Interpersonal Interactions with Providers
nated because they did not meet the a priori selection
criteria. Studies were excluded if the emphasis was not The predictor domain most strongly associated with
primarily on patient satisfaction with ED care (n ⫽ 16), global ED satisfaction (i.e., overall satisfaction, likeli-
the study was conducted outside of the United States (n hood to recommend, or intention to return) across 10 of
⫽ 21), the article did not have sufficient description of the 13 multivariate studies was patient satisfaction with
methods (n ⫽ 10), and if a quantitative research design interpersonal interactions with ED providers (29,30,32–
was not used (n ⫽ 10). A list of the references for the 34,37– 40). Although each of the remaining predictor
excluded studies is available by request from the senior domains has been found to be the strongest predictor of
author. patient satisfaction in at least one study, none has been
The 50 articles retained were observational, univariate replicated as frequently as provider interactions, and
(n ⫽ 25), (4 –28); observational, multivariate (n ⫽ 13), none has been found to be the strongest predictor in the
(29 – 41); and interventional (n ⫽ 12), (42–53). Tables 1, most methodologically sound studies (i.e., those with an
2, and 3 summarize the methodological dissection and average rating of 2.0 or better).
conclusions derived from the multivariate studies, and In one of the largest studies of ED patient satisfaction
Tables 4, 5, and 6 summarize the interventional studies. conducted to date, Hall and Press surveyed over 3000
Tables describing the 25 observational, univariate studies patients from 23 EDs across the United States (34). A
were not provided to conserve space. However, where factor analysis of survey results yielded four scales re-
appropriate, we used results from the univariate studies flecting satisfaction with nursing care, physician care,
to expand on the conclusions generated from our review waiting and convenience, and diagnostic tests. The nurs-
of the multivariate and interventional studies. ing and physician scales primarily consisted of items
Emergency Department Patient Satisfaction 19

Table 4. Interventional Studies

Study ED type Volume Intervention studied Sample size (% of Assessment

eligible pts) method and time
post-ED visit

Burstin et al. (42) Academic; urban; NA Feedback of quality Pre: 1386 (80%) Post: Tel., 7–12 days
non-profit improvement 2326 (80%)
data and setting
of benchmarks
Corbett et al. (43) Academic 60,000/yr Videotape providing Control: 100 (80.0%) Tel., 7 days
information about Intervention: 98
ED (77.2%)
Debehnke and Decker Academic; level 1 48,000/yr MD/RN patient care Pre: 501 (NA) Post: 454 Mail, 3–5 days
(44) trauma center teams (NA)
Krishel and Baraff (45) Academic NA Informational Control: 100 (NA) On-site survey
handout about Intervention: 100 (NA)
ED
Mayer et al. (46) Academic; Level 60,000/yr Customer service NA Tel., ?? days
1 trauma training
center
Mazor et al. (47) Pediatric; urban 40,000/yr 10-week medical Pre: 85 (90%) Post: 58 On-site survey
Spanish course (90%)
Mowen et al. (48) Urban NA 171 (46%) Mail, 2–5 days
Rydman et al. (49) Urban NA Chest pain Control (regular admit): Interview, prior to
observation unit 52 (NA) Intervention discharge
(obs unit): 52 (NA)
Rydman et al. (50) Urban NA Asthma observation Control (regular admit): Interview, prior to
unit 82 (NA) Intervention discharge
(obs unit): 81 (NA)
Schiermeyer et al. (51) Urban; Level I 60,000/yr MD business cards Pre: 120 (50.0%) Post: Paper and pencil,
trauma center 112 (46.4%) in ED before
leaving
Singer et al. (52) Suburban 50,000/yr TVs in rooms Control: 104 (NA) Paper and pencil,
Intervention: 77 (NA) in ED before
leaving
Spaite et al. (53) Academic; urban; 48,000/yr Process redesign NA 1 ⫽ Tel., ?? days
350 bed 2 ⫽ Mail, ??
hospital; Level days
I trauma

NA ⫽ not available; yr ⫽ year; RN ⫽ nurse; MD ⫽ physician; Tel. ⫽ telephone; ?? ⫽ not reported.

assessing interpersonal and communication skills of the patients to rate their satisfaction with the information
respective professionals. The four factor-derived scales they have been given on such things as the cause of their
were entered with demographics, total length of stay, and presenting complaint, the reasons why they experienced
annual ED volume or census into a multiple regression to delays, and how to manage their problem when they
predict the likelihood of recommending the ED to others. return home. Thompson and colleagues studied more
In the final model, the nursing and physician satisfaction than 1600 ED patients and found that all 11 of their items
scales accounted for the largest portion of variance in assessing expressive quality and information delivery
likelihood of recommending the ED to others. This same exhibited statistically significant relations with satisfac-
trend was noted by Carey and Seibert, who studied tion (27). Most recently, Sun and colleagues studied over
17,271 patients from 235 EDs (33). 2300 patients using 68 potential determinants of satis-
Two aspects of provider interpersonal skills seem to faction with care, including patient characteristics (e.g.,
be imperative: expressive quality and information deliv- demographic variables, diagnoses), process of care mea-
ery. Expressive quality refers to the interpersonal man- sures (e.g., triage status, number of treatments received,
nerisms and perceived humanitarian concern of the pro- final disposition), and patient reported problems (e.g.,
vider. In lay terms, it is often referred to as “bedside not receiving help when needed, not told about potential
manner.” This is assessed by asking patients to rate their delays) (39). The final logistic regression predicting
providers’ friendliness, courtesy, respectfulness, or com- overall satisfaction revealed that 5 of the 10 strongest
passion. Information delivery refers to the amount, qual- determinants were related to information delivery and
ity, and understandability of information given to the adequacy of communication. This finding is further val-
patient during the ED visit. It is assessed by asking idated by the same authors in a follow-up study (40).
20 E. D. Boudreaux and E. L. O’Hea

Table 5. Methodological Review of Interventional Studies*

Internal External
Study representativeness representativeness Study measure Study design Statistics Average

2 2 3 1 3 2.2
Burstin et al. (42) Random selection 5 EDs in Boston Reported psychometrics Pre-post design Linear regression
Good response Used in previous studies controlling for
No comparison b/t potential
enrolled and confounds
non-enrolled
1 1 1 3 3 1.8
Corbett et al. (43) Convenience Single site No psychometrics Randomized, Mann-Whitney U
sample reported controlled trial test
No comparisons
b/t enrolled
versus non-
enrolled
1 1 3 1 3 1.8
Debehnke and Decker Low response rate Single site Widely used Q. with Pre-post design Student’s t-test
(44) No comparisons established
b/t enrolled psychometrics
versus non-
enrolled
1 1 2 2 3 1.8
Krishel and Baraff (45) Convenience Single site No psychometrics Randomization Mann-Whitney U
sample reported procedure
DC’d pts. only Based on other poorly
No comparisons published surveys described
b/t enrolled
versus non-
enrolled
1 1 2 1 1 1.2
Mayer et al. (46) No comparisons Single site Previously validated Pre-post design No statistical
b/t enrolled measure, but poorly analyses
versus non- described reported
enrollded
1 1 2 1 3 1.6
Mazor et al. (47) Pediatric ED Single site No psychometrics Pre-post design Chi-square
9 MDs reported
Based on other
published survey
1 1 3 1 3 1.6
Mowen et al. (48) Selected CCs Single-site Psychometrics reported Random Students t-test
No comparisons Based on other assignment
b/t enrolled and published surveys Manipulation
non-enrolled check showed
failed
1 1 3 3 3 2.2
Rydman et al. (49) No comparisons Single ED Widely used Q. with Randomized Cochrane-Mantel-
b/t enrolled and established controlled trial Haentzel test,
non-enrolled psychometrics Groups similar at Students t-test
onset or Wilcoxin
rank-sum test
used as
appropriate
1 1 3 3 3 2.2
Rydman et al. (50) No comparisons Single ED Widely used Q. with Randomized Students t-test or
b/t enrolled and established controlled trial Wilcoxin rank-
non-enrolled psychometrics Groups similar at sum test used
onset as appropriate
1 1 1 1 1 1.0
Schiermeyer et al. (51) Selected times Single ED No psychometrics Pre-post design Student’s t-test
No comparisons reported used, but data
b/t enrolled are non-
versus non- parametric
enrolled
(Continued)
Emergency Department Patient Satisfaction 21

Table 5. continued

Study Internal External Study measure Study design Statistics Average

representativeness representativeness

1 1 1 3 1 1.4
Singer et al. (52) No comparisons Single ED No psychometrics Randomized Student’s t-test
b/t enrolled and reported controlled trial used, but data
non-enrolled Groups similar at are non-
onset parametric
1 1 1 1 1 1.0
Spaite et al. (53) No description of Single site No psychometrics Pre-post design Only descriptives
sample reported reported
Poor description of
survey

b/t ⫽ between; Q ⫽ questionnaire.

* The following rating system was applied to the studies summarized above: Internal Representativeness (how well the sample
represents the ED population for which the intervention was designed): 3 ⫽ excellent, representativeness demonstrated by comparing
those enrolled to those not enrolled and showing 100% equivalent on variables of importance; 2 ⫽ good, representativeness
demonstrated by comparison b/t those enrolled and those not enrolled and showing equivalence on most but not all variables of
importance; 1 ⫽ poor, questionable representativeness either because of no comparison of those enrolled to those not enrolled or
comparison showing gross dissimilarity on important variables. Generalizability to Other EDs (how well does the sample represent
United States EDs as a whole?): 3 ⫽ excellent, multi-site, national sample; 2 ⫽ good, multi-site, regional sample; 3 ⫽ poor, single-site
sample. Characteristics of Study Measure (how psychometrically sound is the instrument?): 3 ⫽ excellent, acceptable reliability and
validity data reported; questionnaire widely used or used by other investigators independent of the study team; 2 ⫽ good, acceptable
reliability and validity data reported, but questionnaire was specifically constructed for the study (i.e., not previously/widely used); 1 ⫽
poor, unacceptable (or no) reliability and validity data reported. Study Design (how well does the study control for confounding factors?):
3 ⫽ excellent, randomized, controlled, and manipulation shown to be successful; 2 ⫽ good, controlled but not necessarily randomized,
manipulation proven successful; 1 ⫽ poor, non-controlled trial (i.e., pre-post) or controlled trials in which manipulation check failed.
Statistical Analyses (are the statistics appropriate?): 3 ⫽ excellent, appropriate parametric or nonparametric statistics used and effect
sizes reported; 2 ⫽ good, appropriate statistics used, but no effect sizes reported; 1 ⫽ poor, used questionably appropriate statistics.

Furthermore, several studies have shown that lan- so intricately interwoven that patients are unable to ac-
guage barriers can be strong determinants of satisfaction tually tease the two apart. This multi-colinearity could
(5,12,29,47). When expressive quality and information account for the fact that technical skills seem important
delivery are deemed unsatisfactory, or patients experi- in some studies but not in others. Although the literature
ence a language barrier that is not adequately addressed, is not sufficient to determine just how important provider
patients tend to be far less satisfied with their ED visit technical skills are in determining satisfaction, it seems
and report being unlikely to recommend the ED to oth- short sighted to deem them unimportant. We can con-
ers. clude, however, that interpersonal interactions play a
large role in determining whether patients evaluate their
providers as technically competent.
Perceptions of Provider Technical Skills

Two studies have found patients’ perception of their Waiting Times

providers’ technical skills to be the best predictor of
global satisfaction, paramount even to ratings of inter- Although others have found that perceived waiting times
personal interactions with providers (35,38). Rhee and remain statistically and clinically significant predictors in
Bird studied 618 ED patients and found that the technical multivariate models, Bursch and colleagues are the only
skills of the nurse and physician were more important researchers to find that it was more important than inter-
than their bedside manner in determining overall satis- personal interactions with providers (30,31,34,38,41).
faction with care (38). Although there are no published All of the studies that have included a measure of both
studies that have investigated how objective measures of perceived and actual waiting times have found that per-
provider technical skills relate to patient satisfaction, ceived waiting times are much more important in deter-
perceived technical ratings have been shown to correlate mining satisfaction than actual waiting times
very highly with the interpersonal interactions with pro- (13,26,30,31,39,48). If waiting times are longer than
viders in the patient satisfaction literature. This link has what the patient expects or deems appropriate, then dis-
been found in both the general outpatient and ED liter- satisfaction is likely to arise, regardless of the actual time
atures (34,35,54 –59). These two aspects of care may be waited. Although perceived waiting times seem to be
22 E. D. Boudreaux and E. L. O’Hea

Table 6. Conclusions of Interventional Studies

Study Conclusions Intervention domain

targeted*

Burstin et al. (42) Benchmarking and quality improvement efforts were associated Process redesign
with a modest decrease in patient-reported problems (e.g.,
not being told about wait time, not being able to
communicate with family), but was not associated with an
improvement in patient satisfaction.
Corbett et al. (43) A videotape describing how the ED works seen by patients in Information delivery
waiting areas was associated with improved satisfaction with
wait times and reduced anxiety for low to moderately acute
patients.
Debehnke and Decker (44) Using MD/RN patient care team improved patient satisfaction Process redesign
across several domains, including satisfaction with wait to
treatment area, wait to MD assessment, feelings of being
cared for as a person, overall satisfaction, and likelihood to
recommend.
Krishel and Baraff (45) Patients receiving a brochure describing how the ED works Information delivery
rated satisfaction higher across several domains, including
overall satisfaction.
Mayer et al. (46) Customer service training was associated with an increase in Interaction with
patient satisfaction across a range of domains. providers
Mazor et al. (47) A 10-week medical Spanish course for pediatric ED MDs was Interaction with
associated with a decrease in interpreter use and an increase providers
in parent satisfaction with interpersonal skills of providers.
Mowen et al. (48) Providing patients with a standardized verbal estimate of wait Information delivery/Wait
to be seen by the triage nurse may not be enough to impact time perception
patients’ perceptions of wait times.
Rydman et al. (49) Chest pain patients were more satisfied and had fewer Process redesign
problems across several domains if seen in the chest pain
obs unit rather than a standard hospital floor, but less
certainty regarding financial information.
Rydman et al. (50) Asthma patients were more satisfied and had fewer problems Process redesign
with rapid diagnosis and treatment if seen in the asthma obs
unit rather than a standard hospital floor, but less certainty
regarding financial information.
Schiermeyer et al. (51) Handing a patient a business card with MD information upon Information delivery
introduction improved overall satisfaction ratings.
Singer et al. (52) Having a TV in a room was not effective in improving overall Wait time perception
satisfaction for moderately acute patients.
Spaite et al. (53) Process redesign can successfully lead to reduced throughput Process redesign
times and increased patient satisfaction in an Academic ED.
Process redesign included: addition of administrative, clinical,
and ancillary personnel; move to bedside registration;
logistical changes in radiology and laboratory; and re-
arranging bed zones.

MD ⫽ physician; RN ⫽ nurse.
* Intervention domains: interaction with providers, Process redesign, Information delivery, Wait time perceptions.

much stronger predictors of ED patient satisfaction than care precisely because they receive greater interpersonal
actual waiting times, they do not seem as strong as attention from ED providers or get seen faster than those
interpersonal interactions with providers. who are less acute.
Specific predictors that have been included in one or
more multivariate studies but have not demonstrated
Other Variables
statistically significant relations with ED patient satisfac-
Acuity has shown some promise as an important predic- tion in final models are: gender, marital status, insurance
tor of ED patient satisfaction in one study, but this status, presence of pain, presence of chronic illness, type
finding has not been replicated (35). Although it seems of treatment (medical vs. surgical), number of previous
unlikely that acuity is more important than those factors ED visits, length of health plan membership, time/day of
mentioned above, it merits further study, especially as a arrival, diagnosis, disposition, daily ED census, annual
potential moderator of the patient-provider relationship. ED volume, satisfaction with tests, and satisfaction with
More acute patients may be more satisfied with their ED registration.
Emergency Department Patient Satisfaction 23

Interventions to Improve ED Patient Satisfaction Table 7. Suggestions for Improving Interpersonal

Interactions

The results from the studies reviewed above can be used Expressive quality
to guide changes needed to improve ED patient satisfac- Verbal techniques
Introduce oneself by name
tion. Our review findings suggest that enhancing inter- Explain one’s role in the ED
personal skills of providers and altering perceptions of Use reflective listening (i.e., summarizing what the patient
wait times can be targeted as priority areas. However, the has said to demonstrate understanding)
Use empathetic comments such as “I understand” or “I
true test of an intervention is the randomized, controlled see”
trial. Unfortunately, the ED patient satisfaction literature Apologize for waits and delays
is not well represented by such trials. Many of the Apologize for interruptions
Nonverbal techniques
published intervention studies used pre-post designs, and Have good eye contact
those that used randomized, controlled trials suffered Smile (when appropriate)
from significant limitations, such as poor blinding and Adopt a “concerned” and “interested” look that shows
you are listening
inadequately detailed randomization procedures. Conse- Allow the patient to describe their problems without
quently, it is difficult to state with certitude what will undue interruptions
lead to improved patient satisfaction in the ED; however, Information delivery
Use anticipatory guidance (i.e., explain to the patient what
a few research-based suggestions can be made. to expect next)
Provide information about diagnoses and potential causes
of the problem
Explain results of tests and their implications
Information Delivery Provide instructions regarding how they should care for
themselves upon discharge
Explain the purpose of procedures and the potential for pain
Interventions involving improving information delivery and discomfort
have been studied, with the rationale being that better Tailor the content to the intellectual level, medical
sophistication, and language mastery of the patient
informed patients are more likely to be satisfied. Two Foreign language communication
randomized controlled trials have studied the effects of Acquire proficiency in languages most common to the
providing information on how the ED functions (e.g., region
Use professional interpreters
role of triage, use of consultants), with one using a
printed brochure given to the patient after triage and the
other using a videotaped message played in the waiting
area (43,45). Both studies found that providing such
information improved patient satisfaction. Using a pre- Spaite and colleagues demonstrated that process redesign
post design, Schiermeyer et al. also showed that even a can successfully lead to reduced throughput times and
simple strategy such as giving patients a business card increased patient satisfaction in an academic ED (53).
upon greeting improved satisfaction (51). However, not Process redesign included: addition of administrative,
all information delivery interventions seem effective. clinical, and ancillary personnel; move to bedside regis-
Mowen and colleagues showed that providing patients tration; logistical changes in radiology and laboratory;
with a standardized verbal estimate of the wait to be seen and re-arranging bed zones. Burstin and colleagues also
did not affect patients’ perceptions of waiting times or showed that providing benchmarks and performance
their satisfaction (48). These studies all suffered from feedback on several clinical indicators was associated
significant methodological weaknesses, including use of with a modest improvement in the number of patient-
patient satisfaction measures without adequate psycho- reported problems (42). However, no changes in patient
metric data, poor description of randomization proce- satisfaction were seen. Debehnke and Decker showed
dures, failure to document treatment integrity valida- that using MD/RN patient care teams rather than tradi-
tions, and inadequate blinding (43,45,48,51). Although tional “zone” or “unspecified” patient assignment im-
their results are suggestive, this area remains open for proved patient satisfaction across several domains, in-
future investigation using methodologically sound, ran- cluding satisfaction with waiting times, feelings of being
domized controlled trials. cared for as a person, overall satisfaction, and likelihood
to recommend (44). Finally, Rydman and colleagues
showed that establishment of observation units for
Process Redesign asthma and chest pain patients were associated with
improvement in patient satisfaction when compared to
Several pre-post studies investigating the effect of pro- patients who were admitted to the hospital in the “tradi-
cess redesign have yielded mostly favorable findings. tional” fashion (49,50).
24 E. D. Boudreaux and E. L. O’Hea

Table 8. Suggestions for Improving Future Research changes can be made to improve satisfaction? A review
Improving sample representation and generalizability
of the multivariate studies clearly revealed that the stron-
Increase response rates by using multiple attempts to obtain gest predictor of ED patient satisfaction is how satisfied
information the patient is with interpersonal interactions with ED
Expand representation of sample (i.e., include admitted
patients)
physicians and nurses. Perceived technical skills and
Report ED’s characteristics, such as volume, location (i.e., perceived waiting times also seem to be important,
urban versus rural), acuity, staffing ratio, presence of though not as important as provider interactions. Unfor-
residency trained physicians.
Use multiple sites
tunately, due to the scarcity of well-designed, random-
Improving reliability and validity of assessment ized, controlled trials, we could derive no definitive
Use reliable and valid measures conclusions about the effectiveness of any interventions.
Report reliability for sample in question
In multivariate research, include a range of predictors such
There are several suggestive possibilities that warrant
as provider interpersonal interaction, perceived technical consideration, however. These include providing printed
skills, perceived wait times, acuity, and facility or videotaped information about how the ED functions,
characteristics
Use multiple criterion variables
using performance improvement methods to target ED
Expand criterion variables to include compliance with process redesign (e.g., moving from traditional to bed-
discharge instructions, actual return to the ED in the side registration), and improving providers’ capacity for
future, continued subscription to a particular healthcare
plan, actual recommendation of the ED to friends and
information delivery and foreign language skills. Below
family, filing of malpractice claims, filing of formal we elaborate on the practical implications of these find-
complaints with hospital administration, or the likelihood ings.
of the patient paying the bill
Improving study design
Most of the multivariate studies and several of the
Use a priori hypotheses generated from theory intervention studies suggest that improving the interper-
Use longitudinal assessment sonal and communication skills of ED providers is the
Use randomized controlled designs to assess the effects of
interventions
intervention most likely to positively impact patient sat-
Improve statistical analyses and interpretations isfaction. Improvements in interpersonal skills should
In multivariate studies, account for multicolinearity target three main areas: expressive quality, information
Use appropriate parametric or non-parametric statistics
based on the sample distribution
delivery, and foreign language communication skills.
Unfortunately, there are few studies that detail exactly
how to accomplish this. Expressive quality is an elusive
trait that is hard to clearly define and even harder to
Interpersonal Interactions change. It involves both the verbal expressions and non-
verbal behaviors of a provider that convey respect, com-
Finally, improving interpersonal and communication passion, and understanding. Although this has not actu-
skills of providers can lead to improved satisfaction. Two ally been studied explicitly in the ED setting, there are
studies demonstrated empirical evidence that enhanced some commonly accepted techniques to improve expres-
provider skills are linked to better patient satisfaction. sive quality (see Table 7). To be effective interperson-
Mayer and colleagues found an 8-h customer service ally, adequate information related to diagnostics, treat-
training program was associated with an increase in ment, and reasons for delays must be given. It is
patient satisfaction across a range of domains, including particularly important to remember that patients do not
ratings reflecting expressive quality, information deliv- simply want to be given a diagnosis; they want to know
ery, and global satisfaction (46). Mazor and colleagues what the diagnosis means and what the provider believes
found that a 10-week Spanish language training program may be the underlying cause of their condition. When
improved satisfaction scores related to expressive quality such cannot be determined during the ED visit, the
and information delivery for Spanish-speaking families possible etiologies can be reviewed briefly and the pa-
(47). Although both studies are suggestive, it should be tient can be encouraged to visit an outpatient physician
noted that they both have methodological problems such for additional work-up. Patients do not like surprises, nor
as use of a pre-post design and small sample size (n ⫽ 9 do they like to be treated in an authoritarian manner.
physicians) (46,47). When feasible, before or during an assessment, the pro-
vider should explain the purpose of the examination, who
will perform the examination, and what will occur. If
DISCUSSION done concisely and concurrent with the assessment, this
should not add substantial time to the process. More
We critically evaluated the literature on ED patient sat- elaborate explanations can be given to patients who are
isfaction in an attempt to answer two questions: 1) what able to understand and when circumstances permit. If
are the strongest predictors of satisfaction? and 2) what family members are present, they too should be provided
Emergency Department Patient Satisfaction 25

information and treated as respectfully as the patient. cations are needed. These studies should focus primarily
This is especially true for families of patients in critical on the effects of enhancing interpersonal skills of staff,
condition, because the family is likely to be very anxious improving information delivery, redesigning ED pro-
and seeking information about the loved one. cesses, and modifying perceptions and expectations of
Information delivery does not necessarily have to wait times.
occur solely in the context of the patient-provider rela-
tionship to improve patient satisfaction. Interventions as
simple as providing patients with a business card upon REFERENCES
introduction, providing a printed brochure, or showing a
videotape have been shown to improve satisfaction 1. DiMatteo MR, DiNicola DD. Achieving patient compliance: the
psychology of the medical practitioner’s role. New York: Perga-
(43,45,51). Such simple strategies can be used to aug- mon; 1982.
ment effective communication by the ED staff. 2. Ley P. Satisfaction, compliance, and communication. Br J Clin
Contrary to popular belief, research repeatedly has Psychol 1982;21:241–54.
3. Rubin HR, Wu AW. Patient satisfaction: its importance and how to
shown that actual waiting times and length of stay are measure it. In: Gitnick G, Rothberg F, Weiner JL, eds. The busi-
relatively unimportant in determining satisfaction. What ness of medicine. New York: Elsevier; 1991:396 – 409.
does seem to be important, however, is the patient’s 4. Bruce TA, Bowman JM, Brown ST. Factors that influence patient
satisfaction in the emergency department. J Nurs Care Qual 1998;
subjective experience of the waiting time. Satisfaction 13:31–7.
seems to hinge not on how long the patient actually waits 5. Carrasquillo O, Orav EJ, Brennan TA, Burstin HR. Impact of
but whether this length is consistent with expectations. If language barriers on patient satisfaction in an emergency depart-
ment. J Gen Intern Med 1999;14:82–7.
delays are longer than what the patient expects or deems 6. Clark CA, Pokorny ME, Brown ST. Consumer satisfaction with
appropriate, then dissatisfaction is likely to arise. These nursing care in a rural community hospital emergency department.
findings have led authors to conclude that EDs should J Nurs Care Qual 1996;10:49 –57.
7. Counselman FL, Graffeo CA, Hill JT. Patient satisfaction with
focus on changing waiting time perceptions rather than physician assistants (PAs) in an ED fast track. Am J Emerg Med
actual waiting times. To date, however, there have been 2000;18:661–5.
no randomized trials testing this hypothesis. One strategy 8. Davis BA, Bush HA, Thomas SW. Measuring consumer satisfac-
tion with emergency department nursing care. J Nurs Sci 1997;2:
that has been suggested involves providing patients with 35– 47.
estimates of anticipated waiting times. However, Mowen 9. Derose KP, Hays RD, McCaffrey DF, Baker DW. Does physician
and colleagues have shown that simply providing pa- gender affect satisfaction of men and women visiting the emer-
gency department? J Gen Intern Med 2001;16:218 –26.
tients a standard, verbal estimate, during the triage as- 10. Durston W, Carl ML, Guerra W. Patient satisfaction and diagnostic
sessment, of how long they will wait to be placed in a accuracy with ultrasound by emergency physicians. Am J Emerg
bed is insufficient to improve satisfaction (48). Providing Med 1999;17:642– 6.
11. Fosnocht DE, Swanson ER, Bossart P. Patient expectations for
more detailed or tailored estimates has not yet been pain medication delivery. Am J Emerg Med 2001;19:399 – 402.
studied. Finally, some have suggested making the wait 12. Hayes RP, Baker DW. Methodological problems in comparing
more pleasant or engaging by providing access to a English-speaking and Spanish-speaking patients’ satisfaction with
interpersonal aspects of care. Med Care 1998;36:230 – 6.
television or reading material, allowing family to be 13. Hedges JR, Trout A, Magnusson AR. Satisfied patients exiting the
present, using greeters or patient advocates, and having emergency department (SPEED) study. Acad Emerg Med 2002;9:
the patient complete self-reported medical and historical 15–21.
14. Hostutler JJ, Taft SH, Snyder C. Patient needs in the emergency
assessment information might be useful. This also may department: nurses’ and patients’ perceptions. J Nurs Adm 1999;
not be sufficient, as Singer and colleagues showed that 29:43–50.
simply having access to a cable TV was not sufficient to 15. Kelly AM. Patient satisfaction with pain management does not
correlate with initial or discharge VAS pain score, verbal pain
improve satisfaction (52). rating at discharge, or change in VAS score in the emergency
Although the existing research has enhanced our un- department. J Emerg Med 2000;19:113– 6.
derstanding of ED patient satisfaction, several noticeable 16. Mack JL, File KM, Horwitz JE, Prince RA. Factors associated with
emergency room choice among Medicare patients. J Ambul Care
deficits and limitations characterize this literature. Table Mark 1995;6:45–58.
8 provides recommendations for improving the method- 17. Magaret ND, Clark TA, Warden CR, Magnusson AR, Hedges JR.
ological quality of the research in this area. For the field Patient satisfaction in the emergency department–a survey of pe-
diatric patients and their parents. Acad Emerg Med 2002;9:1379 –
to become more sophisticated and to move into the next 88.
phase of development, we must design and use psycho- 18. McMillan JR, Younger MS, DeWine LC. Satisfaction with hospi-
metrically sound assessment instruments to assess pa- tal emergency department as a function of patient triage. Health
Care Manage Rev 1986;11:21–7.
tient satisfaction, and they must be made readily avail- 19. Nerney MP, Chin MH, Jin L, et al. Factors associated with older
able to researchers. Furthermore, our research needs to patients’ satisfaction with care in an inner-city emergency depart-
move away from simple exploratory models to incorpo- ment. Ann Emerg Med 2001;38:140 –5.
20. Rhee KJ, Allen RA, Bird J. Telephone vs mail response to an
rating theory and using a priori hypotheses. Randomized emergency department patient satisfaction survey. Acad Emerg
controlled trials that test the effects of specific modifi- Med 1998;5:1121–3.
26 E. D. Boudreaux and E. L. O’Hea

21. Rhee KJ, Dermyer AL. Patient satisfaction with a nurse practitio- 40. Sun BC, Adams JG, Burstin HR. Validating a model of patient
ner in a university emergency service. Ann Emerg Med 1995;26: satisfaction with emergency care. Ann Emerg Med 2001;38:527–
130 –2. 32.
22. Richards CR, Richell-Herren K, Mackway-Jones K. Emergency 41. Yarnold PR, Michelson EA, Thompson DA, Adams SL. Predicting
management of chest pain: patient satisfaction with an emergency patient satisfaction: a study of two emergency departments. J
department based six hour rule out myocardial infarction protocol. Behav Med 1998;21:545– 63.
Emerg Med J 2002;19:122–5. 42. Burstin HR, Conn A, Setnik G, et al. Benchmarking and quality
23. Singer AJ, Thode HC. Determination of the minimal clinically improvement: the Harvard emergency department quality study.
significant different on a patient visual analog satisfaction scale. Am J Med 1999;107:437– 49.
Acad Emerg Med 1998;5:1007–11. 43. Corbett SW, White PD, Wittlake WA. Benefits of an informational
24. Stahmer SA, Shofer FS, Marino A, Shepherd S, Abbuhl S. Do videotape for emergency department patients. Am J Emerg Med
quantitative changes in pain intensity correlate with pain relieve 2000;18:67–71.
and satisfaction? Acad Emerg Med 1998;5:851–7. 44. Debehnke D, Decker C. The effects of a physician-nurse patient
25. Sox CM, Burstin HR, Orav EJ, et al. The effect of supervision of care team on patient satisfaction in an academic ED. Am J Emerg
residents on quality of care in five university-affiliated emergency Med 2002;20:267–70.
departments. Acad Med 1998;73:776 – 82. 45. Krishel S, Baraff LJ. Effect of emergency department information
26. Thompson DA, Yarnold PR. Relating patient satisfaction to wait- on patient satisfaction. Ann Emerg Med 1993;22:568 –72.
ing time perceptions and expectations: the disconfirmation para- 46. Mayer TA, Cates RJ, Mastorovich MJ, Royalty DL. Emergency
digm. Acad Emerg Med 1995;2:1057– 62. department patient satisfaction: customer service training improves
27. Thompson DA, Yarnold PR, Williams DR, Adams SL. Effects of patient satisfaction and ratings of physician and nurse skill.
actual waiting time, perceived waiting time, information delivery, J Healthc Manag 1998;43:427– 440.
and expressive quality on patient satisfaction in the emergency 47. Mazor SS, Hampers LC, Chande VT, Krug SE. Teaching Spanish
department. Ann Emerg Med 1996;28:657– 65. to pediatric emergency physicians: effects on patient satisfaction.
28. Wissow LS, Roter D, Bauman LJ, et al. Patient-provider commu- Arch Pediatr Adolesc Med 2002;156:693–5.
nication during the emergency department care of children with 48. Mowen JC, Licata JW, McPhail J. Waiting in the emergency room:
asthma. Med Care 1998;36:1439 –50. how to improve patient satisfaction. J Health Care Mark 1993;13:
29. Baker DW, Hayes R, Fortier JP. Interpreter use and satisfaction 26 –33.
with interpersonal aspects of care for Spanish-speaking patients. 49. Rydman RJ, Roberts RR, Albrecht GL, Zalenski RJ, McDermott
Med Care 1998;36:1461–70. M. Patient satisfaction with an emergency department asthma
30. Boudreaux ED, Ary RD, Mandry CV, McCabe B. Determinants of observation unit. Acad Emerg Med 1999;6:178 – 83.
patient satisfaction in a large, municipal ED: the role of demo- 50. Rydman RJ, Zalenski RJ, Roberts RR, et al. Patient satisfaction
graphic variables, visit characteristics, and patient perceptions. with an emergency department chest pain observation unit. Ann
Am J Emerg Med 2000;18:394 – 400. Emerg Med 1997;29:109 –15.
31. Bursch B, Beezy J, Shaw R. Emergency department satisfaction: 51. Schiermeyer RP, Tayal V, Butzin CA. Physician business cards
what matters most? Ann Emerg Med 1993;22:586 –91. enhance patient satisfaction. Am J Emerg Med 1994;12:125– 6.
32. Campanella HC, Campanella PM, Grayson K. Factors affecting 52. Singer AJ, Sanders BT, Kowalska A, Stark MJ, Mohammad M,
department of defense patient satisfaction in a military emergency Brogan GX. The effect of introducing bedside TV sets on patient
department. Mil Med 2000;165:396 – 402. satisfaction in the ED. Am J Emerg Med 2000;18:119 –20.
33. Carey RG, Seibert JH. A patient survey system to measure quality 53. Spaite DW, Bartholomeaux F, Guiste J, et al. Rapid process
improvement: questionnaire reliability and validity. Med Care redesign in a university-based emergency department: decreasing
1993;31:834 – 45. waiting time intervals and improving patient satisfaction. Ann
34. Hall MF, Press I. Keys to patient satisfaction in the emergency Emerg Med 2002;39:168 –77.
department: results of a multiple facility study. Hosp Health Serv 54. Hall JA, Dornan MC. What patients like about their medical care
Adm 1996;41:515–32. and how often they are asked: a meta-analysis of the satisfaction
35. Mack JL, File KM, Horwitz JE, Prince RA. The effect of urgency literature. Soc Sci Med 1988;27:935–9.
on patient satisfaction and future emergency department choice. 55. Hall JA, Roter DL, Katz NR. Meta-analysis of correlates of pro-
Health Care Manage Rev 1995;20:7–15. vider behavior in medical encounters. Med Care 1988;26:657–75.
36. Raper JL. A cognitive approach to patient satisfaction with emer- 56. Hall JA, Roter DL, Milburn MA. Illness and satisfaction with
gency department nursing care. J Nurs Care Qual 1996;10:48 –58. medical care. Curr Dir Psych Sci 1999;8:96 –9.
37. Raper J, Davis BA, Scott L. Patient satisfaction with emergency 57. Bopp KD. How patients evaluate the quality of ambulatory med-
department triage nursing care: a multicenter study. J Nurs Care ical encounters: a marketing perspective. J Health Care Mark
Qual 1999;13:11–24. 1990;10:6 –15.
38. Rhee KJ, Bird J. Perceptions and satisfaction with emergency 58. Bowers MR, Swan JE, Koehler WF. What attributes determine
department care. J Emerg Med 1996;14:679 – 83. quality and satisfaction with health care delivery? Health Care
39. Sun BC, Adams J, Orav JE, Rucker DW, Brennan TA, Burstin HR. Manage Rev 1994;19:49 –55.
Determinants of patient satisfaction and willingness to return with 59. Donabedian A. The quality of care: how can it be assessed? JAMA
emergency care. Ann Emerg Med 2000;35:426 –34. 1988;260:1743– 8.