Patient Education and Counseling 86 (2012) 297315

Contents lists available at ScienceDirect

Patient Education and Counseling

journal homepage: www.elsevier.com/locate/pateducou

Review

Association between nonverbal communication during clinical interactions

and outcomes: A systematic review and meta-analysis
Stephen G. Henry a,b,*, Andrea Fuhrel-Forbis b,c, Mary A.M. Rogers b, Susan Eggly d
a
Department of Veterans Affairs Ann Arbor Healthcare System, Ann Arbor, USA
b
Department of Internal Medicine, University of Michigan, Ann Arbor, USA
c
Center for Bioethics and Social Sciences in Medicine, University of Michigan, Ann Arbor, USA
d
Department of Oncology, Wayne State University/Karmanos Cancer Institute, Detroit, USA

A R T I C L E I N F O A B S T R A C T

Article history: Objective: To conduct a systematic review and meta-analysis of studies reporting associations between
Received 25 February 2011 patients and clinicians nonverbal communication during real clinical interactions and clinically
Received in revised form 21 June 2011 relevant outcomes.
Accepted 5 July 2011
Methods: We searched 10 electronic databases, reference lists, and expert contacts for English-language
studies examining associations between nonverbal communication measured through direct observa-
Keywords: tion and either clinician or patient outcomes in adults. Data were systematically extracted and random
Nonverbal communication
effects meta-analyses were performed.
Systematic review
Meta-analysis
Results: 26 observational studies met inclusion criteria. Meta-analysis was performed for patient
Patientclinician communication satisfaction, which was assessed in 65% of studies. Mental and physical health status were evaluated in
Clinically relevant outcomes 23% and 19% of included studies, respectively. Both clinician warmth and clinician listening were
Direct observation associated with greater patient satisfaction (p < 0.001 both). Physician negativity was not related to
patient satisfaction (p = 0.505), but greater nurse negativity was associated with less patient satisfaction
(p < 0.001). Substantial differences in study design and nonverbal measures existed across studies.
Conclusion: Greater clinician warmth, less nurse negativity, and greater clinician listening were
associated with greater patient satisfaction. Additional studies are needed to evaluate the impact of
nonverbal communication on patients mental and physical health.
Practice implications: Communication-based interventions that target clinician warmth and listening
and nurse negativity may lead to greater patient satisfaction.
Published by Elsevier Ireland Ltd.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298
2. Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298
2.1. Data source and searches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298
2.2. Study selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298
2.3. Data extraction and quality assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299
2.4. Data synthesis and analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299
3. Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299
4. Discussion and conclusion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 308
4.1. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 308
4.2. Conclusion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 309
4.3. Practice implications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 309
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310
Appendix A. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310
Appendix B. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311
Appendix C. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313

* Corresponding author at: Robert Wood Johnson Foundation Clinical Scholars Program, University of Michigan, 6312 Medical Science Building 1, 1150 W. Medical Center
Drive, Ann Arbor, MI 48109, USA. Tel.: +1 734 647 4844; fax: +1 734 647 3301.
E-mail address: henrstep@umich.edu (S.G. Henry).

0738-3991/$ see front matter . Published by Elsevier Ireland Ltd.

doi:10.1016/j.pec.2011.07.006
298 S.G. Henry et al. / Patient Education and Counseling 86 (2012) 297315
1. Introduction
Clinicians increasingly use telephone and electronic communi-
cation to provide care, but communication during face-to-face
interactions remains a central component of patient-centered care
and has been shown to inuence clinically relevant outcomes in a
variety of settings [13]. Most research on the links between
communication and outcomes has focused on verbal communica-
tion. However, a large body of research has shown that nonverbal
communication also plays a central role in face-to-face human
interactions and is especially important for conveying emotional
and relational information [4,5]. Largely because of this research,
nonverbal communication is considered an important component
of face-to-face clinical interactions [6,7], but whether nonverbal
communication inuences outcomes subsequent to interactions
remains poorly understood [8]. Better understanding of the
association between nonverbal communication and clinically
relevant outcomes (e.g., patient satisfaction, adherence, and health
status) is important for designing interventions to improve
patients and clinicians communication skills and to promote
patient-centered care [8].
Unfortunately, evaluating the links between nonverbal
communication and clinically relevant outcomes is difcult for
several reasons. The criterion standard for evaluating nonverbal
communication is direct observation of clinical interactions (e.g.,
video recordings or real time observation), which is often
intrusive and resource intensive [9,10]. Potential causal path-
ways linking nonverbal communication and clinically relevant
outcomes are also difcult to determine. For example, patients
nonverbal communication can both inuence and be inuenced
by the patients health status, perceptions of the clinician, and the
clinicians communication behaviors [11]. Similarly, a clinicians
nonverbal communication may reect not only a patients health
and the topic being discussed, but also the clinicians perceptions
of the patient and the patients communication behaviors [12
15]. Therefore, associations between nonverbal communication
and clinically relevant outcomes may reect confounding by
participants characteristics, perceptions, or contextual factors.
Studying nonverbal communication is also difcult because
many aspects of nonverbal communication, such as changes in
voice tone and body language, often take place without
participants explicit awareness [16]. Finally, research studies
often evaluate nonverbal and verbal communication separately,
but verbal and nonverbal communication typically occur
simultaneously and are interpreted together during face-to-face
interactions [11,17].
For all these reasons, studies of nonverbal communication in
clinical interactions have often focused on controlled settings (e.g.,
interactions involving trainees or actors) rather than on natural
settings with real clinicians and patients. Findings from these
studies have generally conrmed that nonverbal communication
plays a central role in conveying emotional and relational
information during clinical interactions [1820]. These studies
have made important contributions to our understanding of
nonverbal communication in clinical settings, but interactions
involving research participants, students, and actors differ in many
important ways from interactions involving real clinicians and
patients [2124]. For example, frequent eye contact between
standardized patients and physician trainees has been shown to
improve standardized patient satisfaction ratings [25], but these
ndings may not generalize to real clinical interactions. Thus
important questions remain about whether and how patients and
clinicians nonverbal communication during clinical interactions is
related to clinically relevant outcomes.
We conducted a systematic review and meta-analysis of studies
reporting associations between patients and/or clinicians non-
verbal communication during real clinical interactions and
clinically relevant outcomes. We had no a priori hypotheses about
expected ndings. Our goals were to evaluate the published
research linking nonverbal communication and these outcomes
and to identify any consistent associations with relevance for
clinical practice.
2. Methods
2.1. Data source and searches
We searched the following databases: Ovid MEDLINE, CINAHL,
PsycINFO, ISI Web of Science, SCOPUS, Anthropology Plus,
Communication & Mass Media Complete, EMBASE, ProQuest
Dissertations & Theses, and ERIC. A review protocol was not
specied in advance. Working with experienced research librar-
ians, we compiled an exhaustive list of nonverbal communication
terms from the existing literature (e.g., nonverbal communication,
facial expression) and cross-referenced them with terms indicating
clinically relevant outcomes (e.g., adherence, satisfaction). All
studies indexed in searched databases as of June 10, 2010 were
potentially eligible. Complete search strategies for Ovid MEDLINE
and EMBASE are provided in Appendices A and B. Additional
potentially eligible articles were identied by manual literature
searches, by examining article reference lists, and by contacting
experts in the eld of nonverbal communication in medicine.
2.2. Study selection
We included any experimental or observational study of
interactions between adult patients and clinicians that examined
associations between nonverbal communication measured
through direct observation and subsequent clinically relevant
outcomes. Nonverbal communication measures included but were
not limited to facial expression, gaze or eye contact, body language
or gestures, touch, laughter, ratings of voice tone (made from audio
recordings that were digitally ltered to obscure verbal content
[26]), and global ratings of clinician or patient affect (made from
video recordings either with or without audio) [4,11]. Affect
ratings are commonly used in nonverbal communication research
because they evaluate emotional state, which is conveyed mostly
nonverbally [17,18]. We dened clinician as any health profes-
sional, including but not limited to physicians, nurses, psychol-
ogists, physical therapists, and psychotherapists, interacting with a
patient to address a physical or mental health problem. We
included only studies of interactions involving real patients and
clinicians that took place for the purpose of managing actual health
problems.
Clinically relevant outcomes included but were not limited to
any of the following: clinician or patient satisfaction, patient
adherence, patient mental or physical health, patient understand-
ing of clinicians recommendations, patient health care utilization,
and clinician malpractice history. Although satisfaction may not be
considered clinically relevant, we included it for three reasons:
satisfaction has been shown to predict other clinical outcomes
[2729], many regulatory agencies use patient satisfaction as a
measure of health care quality [30,31], and satisfaction is
commonly used as an outcome in research involving clinician
patient communication. Complete inclusion and exclusion criteria
are available in Table 1. Demographic factors such as race, age, and
sex have been shown to moderate communication during clinical
interactions [3235]. We did not include these factors as part of
nonverbal communication for the purposes of this review, but we
did not exclude studies that evaluated these moderators if they
also reported associations between nonverbal communication and
outcomes.
 S.G. Henry et al. / Patient Education and Counseling 86 (2012) 297315 299

Table 1
Inclusion and exclusion criteria for study selection.
Inclusion criteria
Independent variables include nonverbal behaviors measured through direct observation (real-time observation, video recording, or audio recording) of actual
clinical interactions between adult patients and clinicians
Dependent variables include measures of clinically relevant outcomes, including patient or clinician satisfaction
Study reported in Englisha
Exclusion criteria
Studies of nonverbal behavior during interactions conducted for research rather than therapeutic purposes
Studies of composite behavior measures that combined verbal and nonverbal behaviorsb
Studies evaluating nonverbal behaviors for purposes other than communicationc
Studies evaluating measures that could not be observed by interaction participantsd
Studies involving actorse
Studies of gender, age, or race as moderators of communication
Studies of communication involving nonverbal patientsf
Studies with fewer than 20 interactions
a
Studies in languages other than English always included English titles and abstracts, so in order to estimate the number of potentially eligible studies published in
languages other than English, language exclusion criteria were applied at the stage of full article review rather than during the initial search.
b
If studies provided sufcient detail to evaluate the association between specic nonverbal behaviors used in the composite measure and outcomes, those specic
nonverbal behaviors, but not the results from the overall composite measure, were included in the review.
c
For example, studies of therapeutic touch, involuntary movements, vocal tics, pre-post studies of reconstructive procedures, or speech therapy interventions.
d
For example, studies of skin conductance or other physiologic measures.
e
Simulated, standardized, or analog patients.
f
For example, patients in comas or with dementia.

2.3. Data extraction and quality assessment
The primary author reviewed all the article titles to exclude
those that were clearly ineligible, and then reviewed the abstracts
of the remaining articles to identify studies that were potentially
eligible. These potentially eligible articles were reviewed in full by
two authors (SGH and AFF) using an abstraction instrument
developed by the authors (Appendix C). Disagreements about
study inclusion were resolved by discussion with the senior author
(SE). In some cases, we emailed study authors for clarication
about study design. Seven of the eight authors we contacted
responded. Multiple studies published in a single article were
included separately, as were studies reporting different associa-
tions using the same data.
As suggested by the Cochrane Collaboration [36] and PRISMA
guidelines [37], we evaluated each studys risks of bias by examining
specic study components relevant to this topic and constructing a
risk of bias table. We evaluated major sources of bias in
observational research, such as participant selection bias and
information bias [38]. Several common sources of bias were not
applicable to this review because no included studies involved
interventions. Potential sources of bias or error relevant to nonverbal
communication research are low inter-rater reliability and failing to
account for verbal content [39,40]. Verbal and nonverbal communi-
cation happen simultaneously during interactions, so the interpre-
tation of specic nonverbal communication behaviors often
depends on both verbal communication behaviors and the topic
under discussion. For example, a clinicians smile might convey
encouragement and support in the context of a patients successful
weight loss, but sadness and empathy in the context of discussing
unfavorable test results with a patient. No single criterion standard
exists for controlling for verbal content and context [9,39]. Two
common methods are including verbal communication variables as
covariates in analyses and restricting the study sample to uniform
clinical scenarios so that verbal content and the relationship
between verbal and nonverbal communication are less likely to
vary between interactions [41]. Potential sources of error and bias
were assessed for the purposes of this review and are not necessarily
relevant to the original goals of the included studies.
2.4. Data synthesis and analysis
For included studies, we evaluated all reported associations
between nonverbal communication and clinically relevant out-
comes; we evaluated summary or composite measures of
nonverbal communication when studies used such measures for
their primary analysis. We classied studies according to the types
of nonverbal communication and outcomes assessed.
For those studies in which the communication variable,
outcome variable and estimate of effect were similar for at least
3 studies, we conducted random effects meta-analyses [42]. Three
measures of nonverbal communication were evaluated: (1)
clinician warmth, measured by ratings of warmth, caring, or
sensitivity; (2) clinician negativity, measured by ratings of
negativity, anger, hostility, or being argumentative; and (3)
clinician listening, measured by ratings of listening or interest.
The unifying estimate of effect in these studies was the correlation
coefcient (r); the Fisher transformation was used to obtain a
mean and standard error of the normal distribution prior to
pooling in meta-analysis. Forest plots were graphed to illustrate
results across studies. To assess heterogeneity we used Cochrans Q
test, between-study variation (t2), and I2 (measuring the degree of
consistency of study results). For studies that reported estimates of
effect other than r, we converted these estimates to r and included
them in meta-analyses whenever possible.
3. Results
The database search retrieved 6536 articles. Of these, 6269 were
excluded after reviewing article titles and another 210 were
excluded after reviewing article abstracts. After reviewing the
remaining 51 English-language publications in full, 26 studies met
our inclusion criteria. Fig. 1 summarizes the study selection process.
Included studies (23 cohort studies and 3 case-control studies) are
described in Table 2. Most studies were conducted in the United
States (77%) and in primary care settings (65%). The median sample
size was 82 interactions [mean = 145; range 20749].
Included studies generally measured nonverbal communica-
tion using one of two methods: quantifying specic nonverbal
behaviors or rating nonverbal communication on global affect
scales (e.g., anxiety or warmth). Facial expression was almost
exclusively evaluated with highly specialized coding systems and
so was a notable exception to these two methods. Summaries of
the results from each study, along with the measures of nonverbal
communication and outcomes used in each study, are shown in
Table 3.
Patient satisfaction was by far the most commonly assessed
outcome; 17 of the 26 included studies measured patient
300 S.G. Henry et al. / Patient Education and Counseling 86 (2012) 297315

satisfaction. Mental and physical health status were also common

outcomes, assessed in six and ve studies, respectively. Three of
the ve studies that assessed physical health as an outcome used
patient self-report as the only measure of physical health. Four
cohort studies [4346] assessed outcomes beyond the immediate
post-interaction period, and two studies [47,48] evaluated patients
during serial visits over time. The majority of outcomes, therefore,
were measured by either patient or clinician report immediately
after the interaction.
A meta-analysis was performed for those studies that measured
similar communication and outcome variables. Patient satisfaction
was the only outcome for which we were able to perform meta-
analyses. For the association between clinician warmth and patient
satisfaction, the pooled (mean) estimate of effect was 0.31 (95% CI:
0.230.38; p < 0.001), indicating a signicant increase in patient
satisfaction with greater clinician warmth. Cochrans Q test for
heterogeneity yielded a p value of 0.228 (test statistic = 6.90), with
I2 = 27.5% and between-study variance (t2) of 0.0024 all
suggesting little variability in the effect across studies. Fig. 2
shows the forest plot for clinician warmth. The effect was similar
for physicians and nurses. In the included studies [26,45,49,50],
patients rated their satisfaction with nurses professional manner.
Because an overall measure of patient satisfaction with nurses was
not available from these studies, we conducted a sensitivity
analysis using patient satisfaction with nurses competence. The
results were similar to those above (pooled estimate of
effect = 0.32 (95% CI: 0.210.42; p < 0.001; I2 = 0%), suggesting
that nurses warmth is positively associated with greater patient
satisfaction with the nurse competence.
We were also able to conduct a meta-analysis for the
association between ratings of clinician negativity and patient
satisfaction across studies [26,45,50,51]. The pooled estimate of
Fig. 1. Flow chart of study selection.
effect was 0.17 (95% CI: 0.34 to 0.01; p = 0.06). Fig. 3 shows a

Table 2
Study characteristics (studies are organized in reverse chronological order within study design).

Study Study Setting Participants (clinicians; Type of nonverbal Outcomes

design patients; observed interactions) communication

Gilbert and Hayes [43] Cohort Primary care clinic 31 nurse practitioners; 155 Patient and clinician body Patient satisfaction; patient
patients; 155 visits language adherence; patient mental
health; patient physical health
Haskard et al. [50] Cohort Primary care clinic 81 nursing staff; 235 patients; Patient and clinician global Patient satisfaction; clinician
235 visits affect satisfaction
Haskard et al., Cohort Primary care clinic 51 physicians; 199 patients; 199 Physician tone of voice Patient satisfaction; clinician
study a [26] visits satisfaction; patient
adherence; patient physical
health; patient mental health
Haskard et al., Cohort Primary care clinic 142 nurses and physicians; 272 Patient and clinician tone of Patient satisfaction; clinician
study b [26] patients; 541 interactionsa voice satisfaction
Prkachin et al. [46] Cohort Workers Compensation Physicians and physical Patient body language; patient Patient physical health; patient
clinic therapistsb; 148 patients; 148 facial expression health care utilization
initial workers compensation
claim evaluations
Troisi et al. [56] Cohort Inpatient psychiatric 1 psychiatrist; 28 patients with Patient facial expression Patient mental health
facility schizophrenia; 28 interviews
Merten [48] Cohort Inpatient psychiatric 10 therapists; 10 patients; 119 Patient and clinician facial Patient mental health
facility therapy sessions expression
Ambady et al. [44] Cohort Inpatient medical 11 physical therapists; 48 Clinician global affect; clinician Patient physical health; patient
facility patients; 57 physical therapy body language; clinician gaze; mental health
sessions clinician facial expression
Koss and Rosenthal [64] Cohort Primary care clinic 24 physicians; 48 patients; 48 Patient and clinician global Patient satisfaction
return visits affect; patient and clinician body
language
Van Dulmen et al. [47] Cohort Internal medicine 1 physician; 18 patients; 44 new Clinician global affect; clinician Patient satisfaction
referral clinic patient visits for diabetes gaze
controlc
Hall et al. [55]d Cohort Primary care clinic 119 clinicians; 749 patients; 749 Patient and clinician tone of Patient mental health; patient
visits voice; clinician body language; physical health
clinician gaze
Greene et al. [74] Cohort Primary care clinic 18 physicians; 81 patients; 81 Patient and clinician (shared) Patient satisfaction
new patient visits laughter
 S.G. Henry et al. / Patient Education and Counseling 86 (2012) 297315 301

Table 2 (Continued )

Study Study Setting Participants (clinicians; Type of nonverbal Outcomes

design patients; observed interactions) communication

Hall et al., study a [53] Cohort Primary care clinic 50 internists; 97 patients; 97 Clinician gaze; patient and Patient satisfaction
return visits clinician tone of voice
Hall et al., study b [53] Cohort Primary care clinic 127 physicians; 537 patients; Patient and clinician tone of Patient satisfaction
524 return visits voice
von Baeyer [75] Cohort Pain clinic 1 anesthesiologist; 58 patients; Patient global affect; patient Patient health care utilization
58 new patient visits body language
Bensing [52] Cohort Primary care clinic 27 general practitioners; 103 Clinician gaze Patient satisfaction
patients; 103 visits for
hypertension
Larsen and Smith [73] Cohort Primary care clinic 15 physicians; 34 patients; 34 Patient and clinician body Patient satisfaction; patient
new patient visits language; patient and clinician knowledge
eye contact; patient and
clinician touch
Street and Buller [51] Cohort Primary care clinic 10 medical residents; 38 Patient and clinician body Patient satisfaction
patients; 38 visits language; patient and clinician
eye contact; patient and
clinician touch
Comstock et al. [49] Cohort Primary care clinic 15 medical residents; 150 Physician body language Patient satisfaction
patients; 150 visits
Hall et al. [45] Cohort Primary care clinic 2 physicians; 50 patients; 50 Patient and clinician tone of Patient satisfaction; patient
visits voice adherence
Smith and Polis [62] Cohort Primary care clinic 29 physicians; 29 patients; 29 Patient and clinician body Patient satisfaction; patient
new patient visits language; clinician gaze knowledge
Weinberger et al. [61] Cohort Primary care clinic 20 medical residents; 82 Patient and clinician body Patient satisfaction; clinician
patients, 82 visits language; clinician touch satisfaction
Kurtz [63] Cohort Primary care clinic 50 physicians; 50 patients; 50 Clinician body language Patient satisfaction
visits
Rasting and Beutel [67] Case- Inpatient psychiatric 2 therapists; 20 patients; 20 Patient and clinician facial Patient mental health
control facility intake interviews expression
Ambady et al. [57] Case- General and orthopedic 57 general or orthopedic Clinician tone of voice Clinician malpractice history
control surgery clinics surgeons; 114 patients; 114
visits
Dalton et al. [54] Case- Emergency department 28 patients; 28 ED visits for Patient facial expression Patient physical health
control chest paine
a
Included 61 primary care physicians and 81 nurses; 272 patients were rated for 272 physician interactions, 269 of which were preceded by nurse interactions.
b
Number of clinicians not reported; all were physicians or physiotherapists employed to perform workers compensation evaluations.
c
Included initial visit and up to 2 subsequent visits for 18 patients.
d
This meta-analysis report pooled results from several studies. Results by individual study are not available.
e
Clinicianpatient interactions were analyzed based on patient expression during interactions; study does not report number of clinicians, clinician characteristics, or
clinician nonverbal communication.

forest plot of the results. For nurses, there was consistency across
studies; the greater the negativity, the less patient satisfaction
with the nurses personal manner (pooled estimate of
effect = 0.35, 95% CI: 0.53 to 0.17, p < 0.001). For physicians,
however, there was little consistency across studies (p = 0.002
from Cochrans Q test for heterogeneity, I2 = 79.7%). The pooled
estimate of effect for the association between physician negativity
and patient satisfaction was 0.07 (95% CI: 0.29 to 0.14;
p = 0.505). In a sensitivity analysis, we conducted a meta-analysis
for the association between nurse negativity and patient satisfac-
tion with nurses competence. The results were similar to those
above (pooled estimate of effect = 0.18 (95% CI: 0.28 to 0.07;
p = 0.001), suggesting that nurse negativity is inversely associated
with patient satisfaction with nurse competence.
Finally, we conducted a meta-analysis of the association
between ratings of clinician listening and patient satisfaction
across studies [26,49,5153]. The pooled estimate of effect was
0.25 (95% CI: 0.130.37; p < 0.001), suggesting a positive associa-
tion between patient satisfaction and clinician listening. However,
there was considerable variation in the effect across studies, with
Cochrans Q statistic = 46.5 (p < 0.001), I2 = 80.7% and t2 = 0.028.
For other communicationoutcome pairs investigated in the
included studies, differences in study design and measures of
nonverbal communication and outcome precluded meta-analysis.
For example, eight studies tested associations between body
language and patient satisfaction. However, each study used a
different method for measuring body language and these measures
were too dissimilar to combine across studies.
Information about potential sources of bias and error in the
included studies can be found in the risk of bias table (Table 4).
Many authors did not report information about participant refusal
rates or participant selection bias. Most studies used convenience
(rather than systematic or random) sampling. Patient attrition was
rarely reported, but is unlikely to be a problem in cohort studies
that follow patients for only a few hours. Nearly all studies that
used more than one judge to measure nonverbal communication
reported some measure of inter-judge reliability. Only 10 studies
used statistical methods that can control for covariates or
moderators. Many studies also tested multiple communication-
outcome associations without adjusting the threshold for type I
error to account for multiple comparisons. For example, research-
ers in one study found that among patients presenting to the
emergency room with chest pain, 4 of 40 separate measures of
facial expression were signicantly associated with positive
cardiac enzymes [54]. However, researchers in this study used
p < 0.1 as the signicance threshold, so on average 4 of 40
measured associations would meet this threshold by chance alone.
Finally, only 9 studies used at least one of the strategies mentioned
previously to control for verbal content when evaluating associa-
tions between nonverbal communication and outcomes.
Two studies assessed patient-reported health status [55] and
clinician-reported mental health status [56] immediately after
clinical interactions. Although nonverbal communication is
unlikely to immediately change mental or physical health, it
could immediately change patients and clinicians perceptions of
patients health. One casecontrol study evaluated associations
 302
Table 3
Results categorized by category of nonverbal communication.

Nonverbal communication Study Nonverbal measure Observation type Outcome Outcome measure Comment on resultsa

Body language
Physician nonverbal gesturesb Comstock Global ratings Real-time Patient satisfaction Satisfaction assessment Not signicant
et al. [49] observation
Physician nonverbal encouragement Weinberger Global rating; Real-time Patient satisfaction; Patient satisfaction More physician nonverbal encouragement (nods, gestures)
and physical distance between et al. [61] distance observation physician satisfaction assessment; physician associated with higher patient (p < 0.001) and clinician
patient and physician satisfaction assessment (p < 0.001) satisfaction; less physical distance between patient
and physician associated with higher patient satisfaction
(p < 0.001)
Physician attentiveness and Kurtz [63] Mehrabian Video without Patient satisfaction Liebig satisfaction Not signicant
dynamism systemc audio measure
Patient and physician distance, Larsen and Mehrabian Video without Patient satisfaction; Satisfaction assessment; More physician neck relation and backward lean associated
body orientation, forward lean, Smith [73] systemc audio patient understanding knowledge assessment with less patient satisfaction (p < 0.01); more direct physician
backward lean, sideways lean, body orientation associated with more patient satisfaction
and limb position (p = 0.05) and better patient understanding (p < 0.01); more

S.G. Henry et al. / Patient Education and Counseling 86 (2012) 297315

patient hand relaxation associated with less patient
satisfaction and less patient understanding (p < 0.01); more
patient social touch and more patient direct body orientation
associated with less patient understanding (p < 0.01)
Patient and physician gestures Street and Difference in Video with audio Patient satisfaction Satisfaction assessment Not signicant
and body orientation Buller [51] frequenciesd [76]
Nurse practitioner and patient Gilbert and Frequency; Video with audio Patient satisfaction; 3 single-item questions; More frequent nurse practitioner nonverbal behaviors
nonverbal activity (nods, gaze, Hayes [43] coordination patient intent to adhere; SF-12 [77] associated with less patient satisfaction (p = 0.03); more
head shakes, eyebrow movements, between nurse change in status of coordination between patient and nurse practitioner
smiles, direct interpersonal practitioner presenting complaint; associated with worse patient physical (p < 0.01) and mental
orientation, touches, backchannels) and patient change in patient mental (p = 0.03) health.
and nurse practitioner-patient and physical health
coordination
Interactional synchrony in physician Koss and Interactional Video without Patient satisfaction Satisfaction assessment Not signicant
and patient body language Rosenthal [64] synchrony; [78] audio [53]
perceived positivity
Time physician spent within Smith and Percent of time Video with audio Patient satisfaction; Satisfaction assessment; More time spent by physician within 3 feet of patient was
3 feet of patient Polis [62] patient comprehension of knowledge assessment associated with increased patient comprehension of
clinician instructions instructions (r = 0.34, p < 0.05)
Physician smiling and nodding Hall et al. [55] Frequency Video without Patient physical and Self-report of mental Not signicant
audio mental health and physical health
status
Physical therapist nonverbal behaviors Ambady Frequency Video without Percent change in patient Confusion assessment More physical therapist shrugs associated with decreased
(sitting, shrugs, forward leans, et al. [44] audio functional status and [44]; geriatric patient mobility at discharge (r = 0.44, p < 0.001)
and head shakes) patient confusion from depression scale [79];
admission; patient mental mobility assessment
status at discharge scale [52]; physical self-
maintenance scale [80];
will to function
assessment [81]
Patient pain behaviors Prkachin Presence of Real-time Patient return to work Workers compensation Presence of patient guarding behavior associated with patient
et al. [46] behaviors observation status 3 months after claims data; SF-36 [82] delayed return to work (p < 0.01), patient-reported physical
visit; patient days of work disability (p < 0.01), more days of work lost (p < 0.05), and
missed; patient injury higher costs of workers compensation claims (r = 0.12,
cost; patient physical p = 0.08)e
health
Patient pain behaviors and patient von Baeyer [75] Number of Video with audio Patient health care Number of return visits; More patient social behaviors associated with more patient
social behaviorsf behaviors present utilization duration of follow up in return visits (r = 0.32, p < 0.01) and longer patient follow-up in
pain clinic the pain clinic (r = 0.40, p < 0.01)
Eye contact/gaze
Patient-directed gaze Bensing [52] Percent of time Video without Patient satisfaction Satisfaction assessment Greater percent of time physician looking at patient associated
audio [49] with greater patient satisfaction (r = 0.19, p = 0.028)e
Patient-directed gaze van Dulmen Percent of time Video with audio Patient satisfaction Medical interview Not signicant
et al. [47] satisfaction scale [83]
Time physician spent looking at Smith and Percent of time Video with audio Patient satisfaction; Satisfaction assessment; More time physician looking at chart associated with less
medical chart Polis [62] patient understanding knowledge assessment patient satisfaction (r = 0.69, p < 0.001) and lower patient
understanding (r = 0.34, p < 0.05)
Time physician spent looking at Hall et al., Length of time Video without Patient satisfaction Satisfaction assessment Not signicant
medical chart study a [53] audio
Patient-directed gaze, physician- Larsen and Mehrabian Video without Patient satisfaction; Satisfaction assessment; Greater physician-directed gaze by patient associated with
directed gaze Smith [73] systemc audio patient understanding knowledge assessment lower patient satisfaction (p < .01)
Physician and patient gazing away Street and Difference in Video with audio Patient satisfaction Satisfaction Not signicant
from one another Buller, 1987 [51] frequenciesd assessment[76]
Patient-directed gaze Ambady Frequency Video without Percent change in patient Confusion assessment Not signicant
et al. [44] audio functional status and [44]; geriatric
patient confusion from depression scale [79];
admission; patient mental mobility assessment
status at discharge scale [52]; physical self-
maintenance scale [80];
will to function

S.G. Henry et al. / Patient Education and Counseling 86 (2012) 297315

assessment [81]
Time physician spent looking at Hall et al. [55] Length of time Video without Patient mental and Self-report mental and Not signicant
medical chart audio physical health physical health status
Facial expression
Patient pain facial expression Prkachin Facial action Real-time Patient return to work Workers compensation Not signicant
et al. [46] coding system [84] observation status 3 months after claims data; SF-36 [82]
visit; patient days of work
missed; patient injury
cost; patient physical
health
Physical therapist facial Ambady Frequency Video without Percent change in patient Confusion assessment Greater therapist facial expressiveness associated with better
expressiveness (composite et al. [44] audio functional status and [44]; geriatric patient activities of daily living (r = 0.60, p < 0.001) and less
of nods, frowns, and smiles) patient confusion from depression scale [79]; patient confusion (r = 0.41, p < 0.01)
admission; patient mental mobility assessment
status at discharge scale [52]; physical self-
maintenance scale [80];
will to function
assessment [81]
Patient and therapist emotion facial Merten [48] Emotion facial Video without Patient mental health Freiburger symptom list Negative therapist emotions (anger, contempt, disgust)
expression action coding audio [86] associated with higher ratings of therapy success by therapist
system [85] (Spearman rho = 0.81, p < 0.01); negative therapist emotions
combined with patient joy associated with higher ratings of
therapy success by patient and therapist (rho = 0.75, p < 0.05);
difference in patient and therapist emotions associated with
more therapy success (rho = 0.61, p < 0.05)
Patient and therapist emotion facial Rasting and Emotion facial Video without Patient mental health General severity index Dissimilar facial emotions between patient and therapist
expression Beutel [67] action coding audio [87] associated with higher therapy success (r = 0.41, p = 0.035)e
system [85]
Patient prosocial facial expression Troisi et al. [56] Ethological Video without Patient mental health Global assessment of Fewer patient prosocial facial expressions associated with
coding system audio functioning [89]; lower patient global assessment of functioning in bivariate
for interviews [88] Sheehan disability scale analysis (r = 0.45, p = 0.02)
Patient facial expression Dalton Facial action Video without Patient physical health Presence of creatine Patient lowering brow associated with presence of creatine
et al. [54] coding system [84] audio kinase (blood test) kinase (p < 0.05)
Global affect
Physician global affect van Dulmen Global ratings Video with audio Patient satisfaction Medical interview Less irritated and more interested physician affect associated
et al. [47] satisfaction scale [83] with greater patient satisfaction (p < 0.05)
Positivity between physician Koss and Global ratings Video without Patient satisfaction Satisfaction assessment Not signicant
and patient Rosenthal [64] audio [53]

303

Table 3 (Continued )
Nonverbal communication
Patient and nurse global affect
Physical therapist global affect
Patient dramatic behavior and
patient distress
Laughter
Patient and physician shared laughter
Tone of voice
Patient and physician tone of voice
Patient and physician tone of voice
Physician tone of voice
304
Study Nonverbal measure Observation type Outcome Outcome measure Comment on resultsa
Haskard Global ratings Video without Nurse satisfaction; patient Patient Satisfaction Nurses rated more caring/sensitive (r = 0.33, p < 0.01),
et al. [50] audio satisfaction Questionnaire [90]; professional (r = 0.35, p < 0.01), and less rushed (r = 0.23,
clinician satisfaction p < 0.05) associated with higher patient satisfaction; patients
assessment rated more pleasant (r = 0.30, p < 0.01) and involved (r = 0.30,
p < 0.01) associated with higher patient satisfaction; nurses
rated more caring/sensitive (r = 0.25, p < 0.05) and patients
rated as more accommodating (r = 0.24, p < 0.05) associated
with higher nursing satisfaction
Ambady Global ratings Video without Percent change in patient Confusion assessment At discharge: therapist distancing associated with worse patient
et al. [44] audio functional status and [44]; geriatric activities of daily living (r = 0.34, p < 0.01); more therapist
patient confusion from depression scale [79]; professionalism associated with worse patient depression
admission; patient mental mobility assessment (r = 0.35, p < 0.01); greater therapist nervousness associated
status at discharge scale [52]; physical self- with more patient will to function (r = 0.29, p < 0.05).
maintenance scale [80]; At 3 months: more therapist professionalism (r = 0.51,
will to function p < 0.0005) and nervousness (r = 0.52, p < 0.0005) associated
S.G. Henry et al. / Patient Education and Counseling 86 (2012) 297315
assessment [81] with worse patient functional status; more therapist distancing
associated with more patient confusion (r = 0.29, p < 0.05) and
worse patient functional status (r = 0.35, p < 0.05)
von Baeyer [75] Global ratings Video with audio Patient health care Number of return visits; More dramatic patient behavior associated with more return
utilization duration of follow up in visits (r = 0.40, p < 0.01) and longer follow up in pain clinic
pain clinic (r = 0.26, p < 0.05)
Green et al. [74] Frequency Audio Patient satisfaction Satisfaction assessment More shared laughter between patient and physician
[91] associated with greater patient satisfaction (r = 0.34, p < 0.01)
Hall et al., Global ratings Filtered audio Patient satisfaction Satisfaction assessment Female physicians and female patients: friendlier (r = 0.52,
study a [53] p < 0.01) and calmer (r = 0.50, p < 0.01) patient voices
associated with greater patient satisfaction; male physicians
and female patients: friendlier patient voices associated with
greater patient satisfaction (r = 0.44, p < 0.05)
Hall et al., Global ratings Filtered audio Patient satisfaction Satisfaction assessment Male physicians and female patients: less dominant physicians
study b [53] (r = 0.16, p < 0.05) and more interested physicians (r = 0.16,
p < 0.05) and patients (r = 0.16, p < 0.05) associated with
greater patient satisfaction; male physicians and male
patients: friendlier patient voices associated with more patient
satisfaction (r = 0.18, p < 0.05). Female physicians and female
patients: less dominant physician voices associated with
greater patient satisfaction (r = 0.30, p < 0.05); female
physicians and male patients: less interested physician voices
associated with greater patient satisfaction (r = 0.36, p < 0.05)
Haskard et al., Global ratings Filtered audio Patient satisfaction; Adherence More warm/supportive (r = 0.38, p < 0.01) and competent/
study a [26] patient physical and questionnaire [92]; interested (r = 0.36, p < 0.01) physicians associated with
mental health; clinician physician satisfaction greater patient satisfaction; more warm/supportive (r = 0.31,
satisfaction; patient assessment [93]; patient p < 0.05) and competent/interested (r = 0.33, p < 0.05)
adherence; patient pain satisfaction physicians associated with greater adherence; more
questionnaire [90]; SF- enthusiastic physicians associated with greater adherence
36 [82] (r = 0.58, p < 0.001); more warm/supportive physicians
associated with greater physician satisfaction (r = 0.38,
p < 0.01); more warm/supportive (r = 0.29, p < 0.05)and
enthusiastic (r = 0.38, p < 0.01) physicians associated with
worse patient mental health; more hostile physicians
associated with worse patient physical health (r = 0.35,
p < 0.05); more warm/supportive(r = 0.36, p < 0.01) and
competent/interested (r = 0.28, p < 0.05) physicians
associated with less pain; more hostile doctors associated with
greater pain (r = 0.35, p < 0.05).
Patient and physician tone of voice
Patient and clinician tone of voice
Hall et al. [45] Global ratings Filtered audio Patient satisfaction; Satisfaction assessment; Angrier physicians associated with greater patient satisfaction
patient adherence percent of no-shows in (bivariate r = 0.37, p < 0.05; multivariate r = 0.64, p < 0.05);
four months following more anxious physicians associated with greater patient
visit satisfaction (bivariate r = 0.32, p < 0.05)
Haskard et al., Global ratings Filtered audio Patient satisfaction; nurse Patient satisfaction More professional physician voices associated with greater
study b [26] satisfaction questionnaire [90] patient satisfaction (r = 0.25, p < 0.05); more involved nurses
associated with greater nurse satisfaction (r = 0.31, p < 0.01);
more caring/interested physicians associated with lower
patient satisfaction (r = 0.26, p < 0.05), more caring/
interested nurse voices associated with higher patient
satisfaction with nurse personal manner and competence
(r = 0.31, p < 0.01 for both); more negative nurses associated
with less patient satisfaction with nurse personal manner
(r = 0.40, p < 0.001); more positive patients associated with
greater patient satisfaction with physician personal manner
(r = 0.29, p < 0.05) and physician competence (r = 0.26,
p < 0.05); more involved (r = 0.25, p < 0.05) and condent
(r = 0.30, p < 0.05) patients associated with greater satisfaction

S.G. Henry et al. / Patient Education and Counseling 86 (2012) 297315

with nurse personal manner.
Patient and physician tone of voice Hall et al. [55] Global ratings Filtered audio Patient mental and Patient mental and More bored physician voices associated with better patient
physical health physical health status physical health (r = 0.12, p < 0.05). More submissive patients
(studies used different associated with better patient physical (r = 0.17, p < 0.05) and
measures); physician mental (r = 0.19, p < 0.05) health and better physician-reported
report mental health (r = 0.20, p < 0.05). More optimistic patient
voices associated with worse patient-reported physical health
(r = 0.24, p < 0.05).
Surgeon tone of voice Ambady Global ratings Filtered audio Surgeon malpractice Clinician malpractice More dominant (partial r = 0.22, p = 0.02) and less concerned/
et al. [57] history claims data anxious (r = 0.18, p < 0.05) surgeons associated with more prior
malpractice claims.
Touch
Patient and physician social touch Street and Difference Video with audio Patient satisfaction Satisfaction assessment Not signicant
Buller [51] in frequenciesd [76]
Patient and physician touch Larsen and Mehrabian Video without Patient satisfaction; Satisfaction assessment; More physician social touch associated with lower patient
Smith [73] systemc audio patient understanding knowledge assessment satisfaction (p < 0.05); more patients social touch associated
with less patient understanding (p < 0.01)
Physician touching patient Weinberger Global ratings Real-time Patient satisfaction; Patient satisfaction Not signicant
during encounter et al. [61] observation clinician satisfaction assessment; clinician
satisfaction assessment
a
Exact p values and measures of effect size (r) are reported (when available) for all statistically signicant results.
b
Dened only as nonverbal gestures such as eye contact and position to patient, and rated as a single global variable called physical attention.
c
Adapted from Mehrabians coding system for nonverbal behavior [94].
d
Analyzed difference in frequencies between physician and patient behaviors.
e
Reported r value was calculated from a different measure of effect using methods described by Rosenthal and DiMatteo [95].
f
Pain behavior was a composite of frowning, grimacing, and referencing painful body parts. Social behavior was a composite of smiling, nodding, illustrative gestures, and averting gaze.

305
306 S.G. Henry et al. / Patient Education and Counseling 86 (2012) 297315

Percent
Author Year Fisher r (95% CI) Weight

Physician
Hall 1981 0.50 (0.21, 0.78) 6.25
Comstock 1982 0.19 (0.03, 0.35) 16.08
Haskard 2008 study b 0.23 (0.11, 0.35) 24.19
Haskard 2008 study a 0.40 (0.26, 0.54) 19.73
Subtotal (I2 = 54.7%, p = 0.085) 0.31 (0.18, 0.43) 66.25

Nurse

Haskard 2008 study b 0.32

0.32 (0.20,
(0.20, 0
0.44)
.44) 24.03
24.03

Haskard 2009 0.34 (0.12, 0.56) 9.72

Subtotal (I2 = 0.0%, p = 0.863) 0.33 (0.22, 0.43) 33.75

Overall (I2 = 27.5%, p = 0.228) 0.31

0.31 ((0.23,
0.23, 0
0.38)
.38) 100.00

NOTE: Weights are from random effects analysis

0 .2 .4 .6 .8
Correlation coefficient (higher numbers indicate stronger
positive associations)
Fig. 2. Association between clinician warmth and patient satisfaction.

Fisher r (95% CI) Percent

Author Year Weight

Physician
Hall 1981 0.39 (0.10, 0.67) 13.66

St
Streett 1987 0 22 ((-0.55,
-0.22 0 55 0.11)
0 11) 12 12
12.12

Haskard 2008 study b -0.15 (-0.27, -0.03) 19.63

Haskard 2008 study a -0.22 (-0.36, -0.08) 18.98

Subtotal (I2 = 79.7%,

79 7% p = 0
0.002
002) -0 07 (-0
-0.07 29 0
(-0.29, 14)
0.14 64.39
64 39

Nurse
-0.42 (-0.54, -0.30) 19.61
Haskard 2008 study b

Haskard 2009 -0.23 (-0.46, -0.01) 16.00

Subtotal (I2 = 53.8%, p = 0.141) -0.35 (-0.53, -0.17) 35.61

Overall (I2 = 83.1%, p = 0.000) -0.17 (-0.34, 0.01) 100.00

NOTE: Weights are from random effects analysis

-.8 -.6 -.4 -.2 0 .2 .4 .6 .8

Correlation coefficient (higher numbers indicate stronger

positive associations)
Fig. 3. Association between clinician negativity and patient satisfaction.
Table 4
Risk of bias for included studies (studies are organized in reverse chronological order within study design).

Study Study design Clinician Patient Participation rate Attrition rate Statistical analysis Accounts Judges blind to Interrater reliability
selection selection for verbal patient outcome or
strategy strategy communi- research question
cation

Gilbert and Hayes [43] Cohort Systematic Convenience Not reported 45% of clinicians; Multiple linear Yesb Not reported All discrepancies in
99% of patients regression nonverbal behavior
ratings resolved by
principal investigator
Haskard et al. [50] Cohort Systematic Convenience Not reported Not reported Pearson correlation No Not reported Cronbachs alpha = 0.52
for patients, 0.64 for
clinicians
Haskard et al. [26] Cohort Systematic Not reported Not reported Not reported Pearson correlation No Not reported Cronbachs alpha = 0.41
Haskard et al. [26] Cohort Not reported Not reported Not reported Not reported Pearson correlation No Not reported Cronbachs alpha = 0.55
for nursing staff
voice, 0.33 for

S.G. Henry et al. / Patient Education and Counseling 86 (2012) 297315

physician voice
Prkachin et al. [46] Cohort Conveniencea Systematica 43% of patientsa 17%a Multiple logistic Yesb,c Yesd Single rater
(secondary regression
analysis)
Troisi et al. [56] Cohort Convenience Convenience Not reported 0% Multiple linear Yesc Yese Single rater
regression
Merten [48] Cohort Not reported Not reported Not reported Not reported Spearman correlation No Not reported Mean Cohens
kappa = 0.89
Ambady et al. [44] Cohort Not reported Not reported Not reported 23% at 3-month Pearson correlation Yesc Not reported Cronbachs alpha = 0.84
follow-up
Koss and Rosenthal [64] Cohort Convenience Convenience Not reported Not reported Pearson correlation No Not reported r = 0.33 for positivity;
r = 0.26 for synchrony
c
van Dulmen et al. [47] Cohort Convenience Convenience Not applicable for Not applicable t-test Yes Not reported Not reported
clinicians; not for clinicians;
reported for 19% for patients
patients
Hall et al. [55] Cohort Not reported Convenience Not reported Not reported Pearson correlation No Not reported Not reported
Greene et al. [74] Cohort Not reported Systematic 95% of clinicians; Not reported Pearson correlation No Not reported Interrater reliability = 0.79
80% of patients
Hall et al. [53] Cohort Systematic Convenience 80% of cliniciansa; Not reported for ANOVA No Not reported Cronbachs alpha for voice
not reported clinicians; 3% quality = 0.67; for time
for patients of patients looking at chart r = 0.97a
Hall et al. [53] Cohort Not reported Convenience Not reported Not reported for ANOVA No Not reported r = 0.78
clinicians; 2%
for patients
von Baeyer [75] Cohort Convenience Convenience Not applicable Not reported Pearson correlation Yesc Yese Single rater
for clinicians;
97% of patients
Bensing [52] Cohort Not reporteda Not reported Not reported for Not reported Discriminant Yesb,c Not reported r = 0.45a
clinicians; analysis
85% of patientsa
Larsen and Smith [73] Cohort Not reported Convenience Not reported Not reported for t-test No Yese Interrater reliability > 0.9
clinicians; 0%
of patients
Street and Buller [51] Cohort Convenience Convenience Not reported for Not reported Pearson correlation No Not reported Cohens kappa
clinicians; range = 0.620.92
86% patients
Comstock et al. [49] Cohort Convenience Convenience 100% of clinicians; 0% Pearson correlation No Yesd 94% of ratings by
93% of patients 2 raters within
1 point on a 4 point scale

307
308 S.G. Henry et al. / Patient Education and Counseling 86 (2012) 297315

between surgeons tone of voice and their history of malpractice

claims [57].

Cronbachs alpha = 0.42

Interrater reliability

for clinicians and

4. Discussion and conclusion
0.44 for patients

Mean interrater
reliability = 0.82

Median r = 0.54

Not reported
Single rater

Single rater
Single rater 4.1. Discussion

Only 26 studies met inclusion criteria for our review, compared

to the wealth of published studies of nonverbal communication in
laboratory, simulated, and non-clinical settings. In meta-analysis,
patient outcome or
research question

overall ratings of clinician warmth were associated with higher

Judges blind to

patient satisfaction scores, with a small-to-moderate effect size

Not reported

Not reported

Not reported
Not reported

(r = 0.3). This association was relatively consistent across studies

Attrition is not applicable to case-control studies; however, in all 3 case-control studies that met inclusion criteria, cases and controls were selected from the same population.
measuring tone of voice [26,45], global affect [50], and body
Yesd

Yesd

Yesd

language [49]. Ratings of clinician negativity were associated with

lower patient satisfaction, but this result was more consistent for
patientnurse interactions than patientphysician interactions.
communi-
for verbal
Accounts

Finally, ratings of clinician listening or interest were associated

cation

with greater patient satisfaction, though there was signicant

Yesb

Yesc
No

No
No

No
No

heterogeneity across studies.

Meta-analyses for outcomes other than patient satisfaction
were not possible for several reasons. First, relatively few studies
Pearson correlation

Pearson correlation
Statistical analysis

evaluated outcomes other than patient satisfaction. Several studies

bivariate analysis

Multiple logistic
Multiple linear

Multiple linear
regression and

from the Netherlands that evaluated associations between

Discriminant

Unadjusted

nonverbal communication and patient mental health status were

regression

regression

odds ratio
analysis

excluded because they measured nonverbal communication

during research interviews [5860]. The methods in these studies
may be useful for future studies of real clinical interactions.
Second, studies measured many different aspects of nonverbal
communication with a variety of measures that could not be
Not applicablef
Not applicablef

Not applicablef
Attrition rate

Not reported

Not reported
Not reported

Not reported

combined across studies. For example, some studies that evaluated

body language measured distance between clinician and patient
[61,62], while others measured the frequency of nonverbal
gestures [43,44,55]. Although measures of distance and frequency
are both reasonable, they are too dissimilar to combine in meta-
analysis. Outcome measures also varied substantially. For exam-
Participation rate

81% of clinicians;

100% of patients
Not reported for

80% of patientsa

ple, included studies measured several different aspects of patient

96% of patients
Not reported

Not reported

Not reported

Not reported

mental health, such as global functioning [56], therapy success

clinicians;

[48], and mental confusion [44]. While combining some of these

Controlled for verbal content by including verbal content variables in statistical analysis.

different measures was mathematically possible, the combined

results would not be interpretable or clinically meaningful. Finally,
several studies (especially older studies) did not report estimates
of effect size or variance, which are required for meta-analysis.
Not reported

Not reported

Convenience
Convenience
Convenience

Convenience

Systematic

Other than clinician type, too few studies reported results of

selection
strategy
Patient

Controlled for verbal content by studying a uniform clinical scenario.

subgroup analyses to allow for investigation of effect moderators

Data from papers cited in study but not reported in the study itself.

in our meta-analyses.
Studies included in this review illustrate the complicated
potential relationships between nonverbal communication and
Not reported

Not reported
Not reported

Not reported
Convenience

Convenience

outcomes. Some included studies evaluated direct associations

Systematic
selection
Clinician

strategy

between nonverbal communication and patient satisfaction. For

example, increased physician attentiveness [63] or more coordi-
nated movements [64] might increase patient satisfaction.
Judges blinded to study research question.

Research outside of clinical settings has shown that nonverbal

Study design

Case-control
Case-control

Case-control

communication and participants emotional reactions tend to

converge during positive interactions [16,65], so that positive
Judges blinded to patient outcome.
Cohort

Cohort
Cohort

Cohort

nonverbal communication during clinical interactions may be the

result rather than the cause of a satisfying interaction. In contrast,
positive associations between pain behaviors and return to work
[46] are more likely to represent confounding by pain severity,
Rasting and Beutel [67]
Weinberger et al. [61]

motivation, or other factors besides direct causation. Associations

Ambady et al. [57]
Table 4 (Continued )

between outcomes and facial expressions (which are largely

Dalton et al. [54]
Smith et al. [62]
Hall et al. [45]

outside of conscious control) [66], are particularly susceptible to

mutual inuence between clinician and patient [16]. To account for
Kurtz [63]

this mutual inuence, future studies of facial expression might

Study

analyze patients and clinicians expressions at the dyadic rather

d
b
a

e
c

than the individual level [48,67].

 S.G. Henry et al. / Patient Education and Counseling 86 (2012) 297315
Our review has several limitations. We were only able to examine
English-language publications (Fig. 1). We did not include unpub-
lished or gray literature in our review and so did not assess
publication bias. However, publication bias generally leads to an
overestimate of signicance [68]. Our review of published studies
revealed only a few consistent associations across studies, so our
conclusions are unlikely to have been affected by excluding
unpublished research. We did not conduct a comprehensive review
of nonverbal communication research in non-clinical or simulated
settings, but as discussed previously we believe focusing on real
clinical interactions is important in order to evaluate the evidence
that is most directly relevant to real outcomes. Studies of simulated
or contrived clinical interactions can provide important insights into
nonverbal communication, but, by denition, they do not inuence
real clinically relevant outcomes. Finally, despite the many potential
direct and indirect ways nonverbal communication may affect
clinical interactions, our review focused only on clinically relevant
outcomes. We believe this focus is critical because efforts to improve
communication during clinical interactions should be ultimately
directed towards improving these outcomes.
4.2. Conclusion
Despite the many challenges of studying nonverbal communi-
cation during real clinical interactions, the existing literature does
suggest that certain global affect ratings clinician warmth,
negativity, and listening are consistently associated with patient
satisfaction. We identied no specic nonverbal behaviors or
gestures (by patients or clinicians) that were consistently
associated with clinically relevant outcomes, nor did we identify
consistent associations between nonverbal communication and
outcomes other than patient satisfaction.
While our meta-analyses did not include all possible nonverbal
measures, our ndings that some ratings were consistently and
signicantly associated with patient satisfaction represent an
important contribution to the literature. These associations (e.g.,
that greater clinician warmth is associated with greater patient
satisfaction) may seem intuitive or obvious, but it is important to
critically evaluate whether obvious associations are supported
by empirical data. For example, although gaze and eye contact play
an important role in everyday conversations [69], we were unable
to perform a meta-analysis to evaluate whether eye contact is
consistently or signicantly associated with patient satisfaction
during real clinical interactions. The estimates of effect size from
our meta-analyses may also help researchers plan future studies
that investigate the associations between nonverbal communica-
tion and patient satisfaction. As discussed previously, association
does not necessarily imply causation. We believe additional
research on the association between nonverbal communication
during real clinical interactions and outcomes is important both to
conrm our ndings and to elucidate the pathways through which
nonverbal communication and patient satisfaction are related.
Our review suggests that existing research on nonverbal
communication during clinical interactions and clinically relevant
outcomes is characterized by many of the same strengths and
weaknesses that experts in the eld described over 15 years ago
[11]. Therefore, we conclude with recommendations for how
future studies can address these limitations and further advance
our understanding of the links between nonverbal communication
and outcomes.
First, our ndings suggest that future studies of real clinical
interactions should take into account the mutual inuence of
clinicians, patients, and context during clinical interactions
[12,16], and the interrelation of verbal and nonverbal communi-
cation [9,35]. Statistical techniques can be used to account for
mutual inuence [70], the relationship between verbal and
309
nonverbal communication [43,51], and the clustered data struc-
ture common in research on clinicianpatient interactions [71,72].
Another limitation of the current evidence on associations
between nonverbal communication and outcomes is the prepon-
derance of descriptive, correlational, hypothesis-generating studies.
Numerous correlation studies have generated many statistically
signicant associations (e.g., between clinicians neck relaxation and
patient satisfaction [73]) that are difcult to interpret or build upon.
Future studies should test specic hypotheses, which could be based
on clinical experience, theories from communication or social
psychology, or results from prior studies involving students or
actors. Studies that evaluate interventions targeting nonverbal
communication (including clinical trials) could also provide
valuable insight regarding the causal pathways underlying associa-
tions between nonverbal communication and outcomes. To our
knowledge no such studies have been conducted using real clinical
interactions. Studies that test hypotheses should specify their
analysis plan prospectively and, when applicable, adjust for multiple
comparisons to reduce the probability that statistically signicant
associations occur by chance.
The variety of nonverbal measures in our review reects the
variety and complexity of nonverbal communication. However, the
use of many different measures also makes it difcult to compare
results across studies. Nonverbal communication researchers
should consider developing a consistent set of validated measures
for research on nonverbal communication during clinical interac-
tions. Use of similar measures would facilitate comparisons across
studies and may make it easier for researchers to build on each
others ndings to advance understanding of the links between
nonverbal communication and outcomes.
Finally, future studies of nonverbal communication and out-
comes should follow patients over longer periods of time in order to
investigate outcomes other than patient satisfaction. For example,
following patients for a few months could allow researchers to
evaluate whether ratings of clinician warmth are associated with
outcomes such as patient adherence and health care utilization.
Nonverbal communication during clinical interactions is unlikely to
immediately affect patient physical or mental health, but it may lead
to changes in health that are mediated, for example, through patient
satisfaction or adherence. Assessing patient physical and mental
health over longer time periods would allow investigators to
investigate these hypotheses prospectively. These recommenda-
tions may seem daunting, but two recent studies in our review
followed all or nearly all of them [43,46]. We believe the majority of
our recommendations can be implemented without setting
unrealistic expectations for future studies in this eld.
Researchers who study the verbal components of clinician
patient communication in isolation have been likened to singers
who study words without music; [11] that is, they can learn a
great deal but cannot capture the full meaning and impact of
communication during clinical interactions. The variety of studies
in this review reects the many different types of nonverbal
communication (e.g., body language, tone of voice), the wide range
of outcomes they may inuence, and the difculties inherent in
studying nonverbal communication during real clinical interac-
tions. Although our ndings indicate associations between
clinician warmth, negativity, and listening and patient satisfaction,
much additional research is needed to better elucidate the
relationship between nonverbal communication and outcomes.
The results of this review and our recommendations for future
studies provide an important starting point for this work.
4.3. Practice implications
Nonverbal communication is an important component of face-
to-face clinical interactions, especially for conveying relational or
310 S.G. Henry et al. / Patient Education and Counseling 86 (2012) 297315

emotional information. Communication-based interventions or 22 Recurrence/

training programs that improve ratings of clinician warmth, 23 px.fs.
listening, or negativity are likely to positively inuence patient
24 therap* efcacy.mp.
satisfaction measured after the interaction. To better understand
the important clinical question of how nonverbal communication 25 outcome*.mp.
inuences outcomes, clinicians and researchers should work 26 utilization*.mp.
together to conduct hypothesis-driven studies that examine 27 exp Drug Utilization/
nonverbal communication during real clinical interactions and
28 exp Drug Utilization Review/
subsequent outcomes. Studies that account for mutual inuence
and context during interactions, follow patients longitudinally, and 29 exp Utilization Review/
measure outcomes other than patient satisfaction are more likely 30 or/1329
to produce results that can not only improve our understanding of Nonverbal communication
nonverbal communication during clinical interactions, but also
31 exp nonverbal communication/
lead to interventions or training programs that improve clinically
relevant outcomes. 32 non-verbal*.mp.
33 nonverbal*.mp.
Conicts of interest 34 exp kinesics/
35 exp Facial Expression/
The authors have no conicts of interest to declare regarding
this study. 36 paralanguage*.mp.
37 para-language*.mp.
Acknowledgements 38 chronemic*.mp.
39 exp Manual Communication/
The authors wish to thank Deborah L. Lauseng, AMLS and Mark
P. MacEachern, MLIS of the A. Alfred Taubman Health Sciences 40 or/3139
Library, University of Michigan for their assistance developing and Potential nonverbal communication behaviors
conducting database searches for this review. They also wish to 41 Eye Movements/
acknowledge Terrance L. Albrecht, PhD and colleagues at the 42 exp Smell/
University of Michigan for helpful advice on earlier versions of this
43 exp Social Dominance/
manuscript, and Joseph Jasperse for editorial assistance. Dr Henry
is supported by the Department of Veterans Affairs and the Robert 44 exp Posture/
Wood Johnson Foundation Clinical Scholars Program. 45 exp Orientation/
46 Gait/
47 exp Spatial Behavior/

Appendix A. Ovid MEDLINE search strategy* 48 Touch/

49 gaze*.mp.
Clinician 50 physical appearance*.mp.
1 exp Physicians/ 51 visual* dominan* ratio*.mp.
2 exp Medical Staff/ 52 body orientation*.mp.
3 exp Nurses/ 53 physical orientation*.mp.
4 therapist*.mp. 54 backward* lean*.mp.
5 practitioner*.mp. 55 forward* lean*.mp.
6 counselor*.mp. 56 body lean*.mp.
7 exp Patients/ 57 gesture*.mp.
8 clinician*.mp. 58 (nod or nods or nodding).mp.
9 Internship and Residency/ 59 proxemic*.mp.
10 doctor*.mp. 60 vocalic*.mp.
11 exp Professional-Patient Relations/ 61 silence*.mp.
12 or/111 62 voice* tone*.mp.
Outcomes 63 pauses.mp.
13 personal satisfaction/ 64 speak* time*.mp.
14 exp Patient Compliance/ 65 back channel*.mp.
15 Malpractice/ 66 self touch*.mp.
16 satisfaction*.mp. 67 haptics.mp.
17 disabilit*.mp. 68 hand shake*.mp.
18 exp Outcome Assessment (Health Care)/ 69 shake* hand*.mp.
19 Outcome and Process Assessment (Health Care)/ 70 guard*.mp.
20 exp Consumer Satisfaction/ 71 (smile or smiling or smiles).mp.
21 (adherence or adhere or adheres).mp. 72 look* away.mp.
 S.G. Henry et al. / Patient Education and Counseling 86 (2012) 297315 311

73 frown*.mp. hierarchically fall beneath it, resulting in a greater number of results.

74 eye contact*.mp. For example, a search of (exp Physicians) will not only include the
75 (bodily position* or body position*).mp. physician concept, but will also include more specic terms, such as
76 physical contact.mp. General Practitioners, Hospitalists, Physicians, Family, and others.
The searches are presented chronologically, and each search
77 pain behavio*.mp.
statement is identied with a number (1 through 91), and the
78 encounter event*.mp.
numbers that appear within search statements correlate to these
79 or/4178 search numbers. Last, the (or/xy) indicates that all search statements
Communication within the numerical range of (xy) are combined with the Boolean
80 exp Communication/ Operator (OR).
81 communica*.mp.
82 non-verbal*.mp. Appendix B. EMBASE search strategy

83 nonverbal*.mp.
Complete nonverbal communication search:
84 exp Nonverbal Communication/
nonverbal* OR nonverbal communication/exp OR nonverbal
85 exp Interpersonal Relations/ communication OR non verbal OR sign language/exp OR manual
86 interpersonal*.mp. communication OR (human relation/exp OR interpersonal com-
87 or/8086 munication/exp OR interpersonal* OR non verbal OR nonverbal*
Potential nonverbal AND communication AND (facial expression/exp OR gesture/exp OR eye movement/exp
88 79 and 87 OR odor/exp OR body posture/exp OR gait/exp OR emotion/exp
OR touch/exp OR gaze/exp OR personal appearance/exp OR
Complete nonverbal communication set
orientation/exp OR paralanguage OR para language OR physical
89 40 or 88
appearance OR backward lean OR forward lean OR body lean OR
Combined: clinician AND outcomes AND complete nonverbal set
nods OR nodding OR proxemic* OR vocalic* OR silence* OR voice
(limited to humans)
tone OR pauses OR speak time OR back channel OR self touch OR
90 12 and 30 and 89 haptics OR hand shake OR shakes hands OR guard OR guarding OR
91 limit 90 to humans smile OR smiles OR smiling OR looks away OR frown* OR eye
contact OR body position OR bodily position OR physical contact
* The (*) at the end of search statements is an Ovid MEDLINE OR pain behavior OR pain behavior OR encounter event OR
truncation symbol. Using (*) to truncate a search term will return all chronemic* OR kinesic*))
conceivable endings for that term. Therefore, a search for commu- Complete outcomes search:
nica* in effect searches for communication, communicating, AND (satisfaction/de OR patient satisfaction/exp OR recurrent
communicate, etc. The (.mp.), which also appears at the end of disease/exp OR recurrence OR treatment outcome/exp OR patient
many search statements, is an Ovid MEDLINE notation that describes compliance/exp)
the elds that the term will be searched as a keyword. These elds Complete clinician search:
generally include title, abstract, subject headings, substance name, AND (medical specialist/exp OR medical expert/exp OR medical
registry words, among others. The (exp) that appears at the beginning staff/exp OR physician/exp OR physician assistant/exp OR psy-
of many search statements is Ovid MEDLINE code for explode, a chotherapist/exp OR resident/exp OR medical personnel/de OR
concept that relates to MEDLINEs Medical Subject Heading (MeSH) nurse/exp OR doctor patient relation/exp OR clinician*)
terminology. The MeSH terminology is structured hierarchically, and Limits:
an exploded (exp) subject heading will pick up all terms that AND [humans]/lim AND ([embase]/lim OR [embase classic]/lim)
312 S.G. Henry et al. / Patient Education and Counseling 86 (2012) 297315

Appendix C. Nonverbal behavior systematic review study

abstraction form

Study Number ___________ Study Year___________ Reviewer ____________

Authors (first 2)______________________________________________________
Journal _______________________________________________________________________

Patient Population:____________________________________ Real patients? Y N stop

number of patients ____________________
controls (if applicable) _____________________

Clinician Population: ___________________________________ Real clinicians? Y N stop

number ____________________

Interaction setting:_____________________________________ Real therapeutic interactions? Y N stop

number____________________

Study design: RTC cohort case-control cross-section other:________________________

Country where study was conducted: USA other:________________________

Is NVB an independent variable? Y N stop

Is dependent variable a patient or physician outcome? Y N stop

Include in systematic review?: N Reason for exclusion __________________________________

Y proceed.

Channels for NVB: real-time video only A+V audio only filtered audio multiple

Type of NVB measure: global ratings + factor analysis counting proportion

global ratings no factor analysis other ________________________

Was primary goal to evaluate associations between NVB and outcomes? Y N other: __________
Attempts to control for content when analyzing NVB? Y N other ____________________________

Notes on study:
 S.G. Henry et al. / Patient Education and Counseling 86 (2012) 297315 313

Study # _______ (if > 10 record only most important ones)

NVB How measured Outcome How measured

#1 #1

#2 #2

#3 #3

#4 #4

#5 #5

#6 #6

#7 #7

#8 #8

#9 #10

#10 #11

Total # of NVBs measured: ______________ Total # of outcomes measured: _____________

Total possible associations: _______________ # of associations reported: _______________

Reports positive & negative results? Y N Controls for multiple associations? Y N

Statistical test of association:_______________________________________

NVB# Out# Sig? Measure NVB Out# Sig? Measure

Y N Y N
Y N Y N
Y N Y N
Y N Y N

References
[1] Haskard-Zolnierek K, DiMatteo MR. Physician communication and patient
adherence to treatment: a meta-analysis. Med Care 2009;47:82634.
[2] Mead N, Bower P. Patient-centred consultations and outcomes in primary
care: a review of the literature. Patient Educ Couns 2002;48:5161.
[3] Williams S, Weinman J, Dale J. Doctorpatient communication and patient
satisfaction: a review. Fam Pract 1998;15:48092.
[4] Knapp ML, Hall JA. Nonverbal communication in human interaction, 7th ed.,
Boston, MA: Wadsworth, Cengage Learning; 2009.
[5] Manusov V, Patterson ML, editors. The SAGE handbook of nonverbal commu-
nication. Thousand Oaks: SAGE Publications; 2006.
[6] Frankel RM, Stein T, Krupat E. The Four Habits approach to effective clinical
communication. Physician education and development. Oakland, CA: Kaiser
Permanente; 2003.
[7] Roter DL, Hall JA. Doctors talking with patients/patients talking with doctors:
improving communication in medical visits, 2nd ed., Westport, CT: Praeger;
2006.
[8] Levinson W, Lesser CS, Epstein RM. Developing physician communication
skills for patient-centered care. Health Affair 2010;29:13108.
[9] Jones SE, LeBaron CD. Research on the relationship between verbal and
nonverbal communication: emerging integrations. J Commun 2002;52:499
521.
[10] Albrecht TL, Penner LA, Cline RJW, Eggly SS, Ruckdeschel JC. Studying the
process of clinical communication: issues of context, concepts, and research
directions. J Health Commun 2009;14:4756.
[11] Hall JA, Harrigan JA, Rosenthal R. Nonverbal behavior in clinician-patient
interaction. Appl Prev Psychol 1995;4:2137.
[12] Street Jr RL, Gordon H, Haidet P. Physicians communication and perceptions of
patients: is it how they look, how they talk, or is it just the doctor? Soc Sci Med
2007;65:58698.
[13] Penner LA, Dovidio JF, West TV, Gaertner SL, Albrecht TL, Dailey RK, et al.
Aversive racism and medical interactions with black patients: a eld study. J
Exp Soc Psychol 2010;46:43640.
[14] Heath C. In: Ekman P, Scherer KR, editors. Body movement and speech in
medical interaction. Cambridge, UK & Paris: Cambridge University Press &
Editions de la Maison des Sciences de lHomme; 1986.
[15] Mast MS. On the importance of nonverbal communication in the physician
patient interaction. Patient Educ Couns 2007;67:3158.
[16] Cappella JN, Schreiber DM. The interaction management function of nonverbal
cues. In: Manusov V, Patterson ML, editors. The SAGE handbook of nonverbal
communication. Thousand Oaks: SAGE Publications; 2006. p. 36179.
[17] Robinson JD. Nonverbal communication and physicianpatient interaction:
review and new directions. In: Manusov V, Patterson ML, editors. The sage
handbook of nonverbal communication. Thousand Oaks, CA: Sage Publica-
tions; 2006. p. 43760.
314 S.G. Henry et al. / Patient Education and Counseling 86 (2012) 297315
[18] Roter DL, Frankel RM, Hall JA, Sluyter D. The expression of emotion through
nonverbal behavior in medical visits: mechanisms and outcomes. J Gen Intern
Med 2006;21:S2834.
[19] Hall JA, Roter DL, Blanch DC, Frankel RM. Observer-rated rapport in interac-
tions between medical students and standardized patients. Patient Educ
Couns 2009;76:3237.
[20] Hall JA, Roter DL, Blanch DC, Frankel RM. Nonverbal sensitivity in medical
students: implications for clinical interactions. J Gen Intern Med 2009;24:
121722.
[21] Henry SG. A Piece of my mind. Playing doctor. J Am Med Assoc 2005;294:
213840.
[22] Wear D, Varley JD. Rituals of verication: the role of simulation in developing
and evaluating empathic communication. Patient Educ Couns 2008;71:
1536.
[23] Malhotra A, Gregory I, Darvill E, Goble E, Pryce-Roberts A, Lundberg K, et al.
Mind the gap: learners perspectives on what they learn in communication
compared to how they and others behave in the real world. Patient Educ Couns
2009;76:38590.
[24] Ram P, van der Vleuten C, Rethans JJ, Grol R. Assessment of practicing family
physicians: comparison of observation in a multiple station examination using
standardized patients with observation of consultations in daily practice. Acad
Med 1999;74:629.
[25] Grifth CH, Wilson JF, Langer S, Haist SA. House staff nonverbal communica-
tion skills and standardized patient satisfaction. J Gen Intern Med 2003;18:
1704.
[26] Haskard KB, Williams SL, DiMatteo MR, Heritage J, Rosenthal R. The providers
voice: patient satisfaction and the content-ltered speech of nurses and
physicians in primary medical care. J Nonverbal Behav 2008;32:120.
[27] Baumann C, Rat AC, Osnowycz G, Mainard D, Cuny C, Guillemin F. Satisfaction
with care after total hip or knee replacement predicts self-perceived health
status after surgery. BMC Musculoskelet Disord 2009;10:150.
[28] Weisman CS, Nathanson CA. Professional satisfaction and client outcomes: a
comparative organizational analysis. Med Care 1985;23:117992.
[29] Street RL, Makoul G, Arora N, Epstein RM. How does communication heal?
Pathways linking clinician-patient communication to health outcomes. Pa-
tient Educ Couns 2009;74:295301.
[30] Performance Measurement. The Joint Commission; Available from: [accessed
20.8.2010].
[31] HCAHPS. Patients perspectives of care survey. Centers for Medicare &
Medicaid Services; Available from: http://www.cms.gov/HospitalQualityInits/
30_HospitalHCAHPS.asp [accessed 20.8.2010].
[32] Cooper LA, Roter DL. Patientprovider communication: the effect of race and
ethnicity on process and outcomes of healthcare. In: Smedley BD, Stith AY,
Nelson AR, editors. Institute of Medicine (U.S.). Committee on Understanding
and Eliminating Racial and Ethnic Disparities in Health Care. Unequal treat-
ment: confronting racial and ethnic disparities in health care. Washington, DC:
National Academy Press; 2003. p. 55293.
[33] Roter DL, Hall JA, Aoki Y. Physician gender effects in medical communication: a
meta-analytic review. J Am Med Assoc 2002;288:75664.
[34] Adelman RD, Greene MG, Ory MG. Communication between older patients and
their physicians. Clin Geriatr Med 2000;16:124. vii.
[35] Dovidio JF, Hebl M, Richeson JA, Shelton JN. Nonverbal communication, race,
and intergroup interaction. In: Manusov V, Patterson ML, editors. The SAGE
handbook of nonverbal communication. Thousand Oaks: SAGE Publications;
2006. p. 481500.
[36] Higgins J, Altman DG. Assessing risk of bias in included studies. In: Higgins J,
Green S, editors. Cochrane handbook for systematic reviews of interventions
version 501: The Cochrane Collaboration. 2008. p. Available from www.co-
chrane-handbook.org [accessed 29.7.2010].
[37] Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gotzsche PC, Ioannidis JPA, et al.
The PRISMA statement for reporting systematic reviews and meta-analyses of
studies that evaluate health care interventions: explanation and elaboration.
Ann Intern Med 2009;151:W6594.
[38] Grimes DA, Schulz KF. Bias causal associations in observational research.
Lancet 2002;359:24852.
[39] Burgoon JK. The relationship of verbal and nonverbal codes. In: Dervin B, Voigt
MJ, editors. Progress in communication sciences. Norwood, NJ: Ablex Publish-
ing; 1985. p. 26398.
[40] LeBaron CJ. Considering the social and material surround: toward microeth-
nographic understandings of nonverbal communication. In: Manusov V,
editor. The sourcebook of nonverbal measures. Mahwah, NJ: Lawrence Erl-
baum Associates; 2005. p. 493506.
[41] Noller P. Standard content methodology: controlling the verbal channel. In:
Manusov V, editor. The sourcebook of nonverbal measures. Mahwah, NJ:
Lawrence Erlbaum Associates; 2005. p. 41730.
[42] DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials
1986;7:17788.
[43] Gilbert DA, Hayes E. Communication and outcomes of visits between older
patients and nurse practitioners. Nurs Res 2009;58:28393.
[44] Ambady N, Koo J, Rosenthal R, Winograd CH. Physical therapists nonverbal
communication predicts geriatric patients health outcomes. Psychol Aging
2002;17:44352.
[45] Hall JA, Roter DL, Rand CS. Communication of affect between patient and
physician. J Health Soc Behav 1981;22:1830.
[46] Prkachin KM, Schultz IZ, Hughes E. Pain behavior and the development of pain-
related disability: the importance of guarding. Clin J Pain 2007;23:2707.
[47] van Dulmen AM, Verhaak PFM, Bilo HJG. Shifts in doctorpatient communi-
cation during a series of outpatient consultations in non-insulin-dependent
diabetes mellitus. Patient Educ Couns 1997;30:22737.
[48] Merten J. Facial microbehavior and the emotional quality of the therapeutic
relationship. Psychother Res 2005;15:32533.
[49] Comstock LM, Hooper EM, Goodwin JM, Goodwin JS. Physician behaviors that
correlate with patient satisfaction. J Med Educ 1982;57:10512.
[50] Haskard K, DiMatteo MR, Heritage J. Affective and instrumental communica-
tion in primary care interactions: predicting the satisfaction of nursing staff
and patients. Health Commun 2009;24:2132.
[51] Street Jr RL, Buller DB. Nonverbal response patterns in physicianpatient
interactions: a functional analysis. J Nonverbal Behav 1987;11:23453.
[52] Bensing J. Doctorpatient communication and the quality of care. Soc Sci Med
1991;32:130110.
[53] Hall JA, Irish JT, Roter DL, Ehrlich CM, Miller LH. Satisfaction, gender, and
communication in medical visits. Med Care 1994;32:121631.
[54] Dalton JA, Brown L, Carlson J, McNutt R, Greer SM. An evaluation of facial
expression displayed by patients with chest pain. Heart Lung 1999;28:168
74.
[55] Hall JA, Roter DL, Milburn MA, Daltroy LH. Patients health as a predictor of
physician and patient behavior in medical visits: a synthesis of four studies.
Med Care 1996;34:120518.
[56] Troisi A, Pompili E, Binello L, Sterpone A. Facial expressivity during the clinical
interview as a predictor of functional disability in schizophrenia: a pilot study.
Progr Neuro-Psychopharmacol Biol Psychiat 2007;31:47581.
[57] Ambady N, LaPlante D, Nguyen T, Rosenthal R, Chaumeton N, Levinson W.
Surgeons tone of voice: a clue to malpractice history. Surgery 2002;132:59.
[58] Bos EH, Bouhuys AL, Geerts E, van Os TW, Ormel J. Stressful life events as a link
between problems in nonverbal communication and recurrence of depression.
J Affect Disorders 2007;97:1619.
[59] Geerts E, Bouhuys N, Van den Hoofdakker RH. Nonverbal attunement between
depressed patients and an interviewer predicts subsequent improvement. J
Affect Disorders 1996;40:1521.
[60] Bouhuys AL, Sam MM. Lack of coordination of nonverbal behaviour between
patients and interviewers as a potential risk factor to depression recurrence:
vulnerability accumulation in depression. J Affect Disorders 2000;57:189
200.
[61] Weinberger M, Greene JY, Mamlin JJ. The impact of clinical encounter events
on patient and physician satisfaction. Soc Sci Med Part E Med Psychol
1981;15:23944.
[62] Smith CK, Polis E, Hadac RR. Characteristics of the initial medical interview
associated with patient satisfaction and understanding. J Fam Pract 1981;12:
2838.
[63] Kurtz S. Physician nonverbal behavior and patient satisfaction in physician
patient interviews. PhD diss. Denver, CO: University of Denver; 1975.
[64] Koss T, Rosenthal R. Interactional synchrony, positivity, and patient satisfac-
tion in the physicianpatient relationship. Med Care 1997;35:115863.
[65] Giles H, Ogay T. Communication accommodation theory. In: Whaley BB,
Santer W, editors. Explaining communication: Contemporary theories and
exemplars. Mahwah, NJ: Lawrence Erlbaum Associates; 2007. p. 32544.
[66] Ekman P, Friesen WV. Unmasking the face: a guide to recognizing emotions
from facial clues. Cambridge, MA: Malor Books; 2003.
[67] Rasting M, Beutel ME. Dyadic affective interactive patterns in the intake
interview as a predictor of outcome. Psychother Res 2005;15:18898.
[68] Dickersin K. Publication bias: recognising the problem, understanding its
origin and scope, and preventing harm. In: Rothstein HR, Sutton AJ, Borenstein
M, editors. Publication bias in meta-analysis: prevention, assessment, and
adjustments. West Sussex, UK: John Wiley & Sons; 2005. p. 1134.
[69] Hall JA, Coats EJ, LeBeau LS. Nonverbal behavior and the vertical dimension of
social relations: a meta-analysis. Psychol Bull 2005;131:898924.
[70] Kenny DA, Kashy DA, Cook WL. Dyadic data analysis. New York: Guilford Press;
2006.
[71] Lipsitz SR, Fitzmaurice GM, Orav EJ, Laird NM. Performance of generalized
estimating equations in practical situations. Biometrics 1994;50:2708.
[72] Raudenbush SW, Bryk AS. Hierarchical linear models: applications and data
analysis methods, 2nd ed., Thousand Oaks: Sage Publications; 2002.
[73] Larsen KM, Smith CK. Assessment of nonverbal communication in the patient
physician Interview. J Fam Pract 1981;12:4818.
[74] Greene MG, Adelman RD, Friedmann E, Charon R. Older patient satisfaction
with communication during an initial medical encounter. Soc Sci Med
1994;38:127988.
[75] von Baeyer CL. Social and pain behavior in the rst 3 min of a pain clinic
medical interview. Pain Clin 1994;7:16977.
[76] Buller MK, Buller DB. Physicians communication style and patient satisfac-
tion. J Health Soc Behav 1987;28:37588.
[77] Ware JE, Kosinski M, Keller SD. How to score version 2 of the SF-12 health
survey. Lincoln, RI: Quality Metric Inc.; 2000.
[78] Bernieri FJ, Reznick JS, Rosenthal R. Synchrony, pseudosynchrony, and dis-
synchrony: measuring the entrainment process in mother infant interactions.
J Pers Soc Psychol 1988;54:24353.
[79] Yesavage JA, Brink TL, Rose TL, Lum O, Huang V, Adey M, et al. Development
and validation of a geriatric depression screening scale: a preliminary report. J
Psychiat Res 1983;17:3749.
[80] Stewart A, Kamberg C. Physical functioning. In: Stewart A, Ware JE, editors.
Measuring functional status and physical well-being: the medical outcomes
study approach. Durham, NC: Duke University Press; 1989. p. 86101.
 S.G. Henry et al. / Patient Education and Counseling 86 (2012) 297315
[81] Hays RD, Sherbourne CD, Mazel RM. The RAND 36-item health survey 1.0.
Health Econ 1993;2:217.
[82] Ware J. SF-36 health survey update. In: Maruish M, editor. Use of psychologi-
cal testing for treatment planning and outcomes assessment: instruments for
adults. 3rd ed., Mahwah, NJ: Erlbaum; 2004. p. 693718.
[83] Wolf MH, Putnam SM, James SA, Stiles WB. The medical interview satisfaction
scale: development of a scale to measure patient perceptions of physician
behavior. J Behav Med 1978;1:391401.
[84] Ekman P, Friesen WV, Hager JC. Facial action coding system: the manual [CD
ROM]. Salt Lake City, Utah: A Human Face; 2002.
[85] Friesen, WV, Ekman, P., Emotion Facial Action Coding System, 1984. Available
at http://www.face-and-emotion.com/dataface/facs/emfacs.jsp. Accessed 22
July 2011.
[86] Fahrenberg J. Die freiburger beschwerdenliste FBL [freiburger symptom list].
Zeitschrift fur Klinische Psychologie 1995;4:79100.
[87] Franke G. SCL-90-R die symptom checklist von derogatis: deutsche version.
Weinheim: Beltz-Verlag; 1995.
315
[88] Troisi A. Ethological research in clinical psychiatry: the study of nonverbal
behavior during interviews. Neurosci Biobehav Rev 1999;23:90513.
[89] Hilsenroth MJ, Ackerman SJ, Blagys MD, Baumann BD, Baity MR, Smith SR, et al.
Reliability and validity of DSM-IV axis V. Am J Psychiat 2000;157:185863.
[90] Marshall GN, Hays RD. The patient satisfaction questionnaire short-form (PSQ-
18): RAND Corporation, 1994 Contract No.: P-7865.
[91] Bertakis KD, Roter D, Putnam SM. The relationship of physician medical
interview style to patient satisfaction. J Fam Pract 1991;32:17581.
[92] Dimatteo MR, Sherbourne CD, Hays RD, Ordway L, Kravitz RL, Mcglynn EA,
et al. Physicians characteristics inuence patients adherence to medical
treatment: results from the Medical Outcomes Study. Health Psychol 1993;
12:93102.
[93] Suchman AL, Roter D, Green M, Lipkin M, Bertakis K, Charon R, et al. Physician
satisfaction with primary-care ofce visits. Med Care 1993;31:108392.
[94] Mehrabian A. Nonverbal communication. Chicago: Aldine-Atherton; 1972.
[95] Rosenthal R, DiMatteo MR. Meta-analysis. Recent developments in quantita-
tive methods for literature reviews. Ann Rev Psychol 2001;52:5982.