Transcript Patient-Physician Communication

Patient-Physician Communication:
Impact on Clinical Outcomes
Meds 1 January 29, 2004
Moira Stewart, Ph.D.
Department of Family Medicine
Canada Research Chair in Primary Health Care
Objectives
1.
to reinforce your learnings on
research methods
2.
to learn about the results of a
research program on patientcentred communication
Tie to Meds 1 Curriculum
1.
follow-up to your Community
Health lectures on research
methods last fall
2.
the results have implications for
your evidence-based clinical
practice from Year 3 on
Outline of the 3 hour session today
Hour 1
Lecture on Evidence for PatientPhysician Communication
25 min
Small Group Discussion on
what are the effective components 20 min
Break
15 min
Outline continued
Hour 2
Mini-lecture on how to evaluate
cohort designs
5 min
Lecture on A Study of the
"Impact of Patient-Centred
Communication on Patient
Outcomes in Family Practice”
- a cohort study
20 min
Small group discussion
- design your own cohort study
20 min
Break
15 min
Outline continued
Hour 3
Mini-lecture on how to evaluate
clinical trial designs
5 min
Lecture on A Study “Innovative
Training to Improve Physician
Communication with Breast
Cancer Patients: Results of a
Randomized Controlled Trial”
15 min
Small group discussion
- design your own clinical trial study
15 min
Closing - a synthesis and summary of
take-home messages with examples of
exam questions
15 min
Hour 1
Lecture on Evidence for PatientPhysician Communication
25 min
Small Group Discussion on
what are the effective components 20 min
Break
15 min
Study Designs

Randomized controlled trials (level I)

Controlled trials without randomizaton (level II - 1)

Cohort or case-control studies (level II - 2)

Comparisons between times or places (level II - 3)

Opinions, descriptive studies (level III)
CTF Recommendation Grades
A.
Good evidence to recommend
B.
Fair evidence to recommend
C.
Conflicting evidence, does not allow for a clear
recommendation and other factors may influence
decision-making
D.
Fair evidence to recommend against
E.
Good evidence to recommend against
I.
Insufficient evidence (quantity and/or quality) to
make a recommendation
Clinical Significance
versus
Statistical Significance
Does Patient-Physician
Communication Matter?
Moira Stewart, Ph.D.
ABMS-ACGME Conference on
Assessment of Physician-Patient
Communication
March 22, 2002
Synthesis of Evidence
What kind of evidence matters?
A-
evidence
 Results with clinical significance
Does the evidence on
patient-physician
communication measure up
to these standards?
YES
The evidence includes

Randomized controlled trials

Clinically significant results on
important outcomes
PHYSICIAN OUTCOMES
PATIENT OUTCOMES
PHYSICIAN OUTCOMES

Malpractice claims

Time

Physician satisfaction
PATIENT OUTCOMES

Satisfaction

Adherence

Health
Stewart. Effective Physician-Patient
Communication and Health Outcomes: A
Review. Canadian Medical Association
Journal. 1995;152(9):1423-1433
Stewart, Brown, Boon et al. Evidence on
Patient-Doctor Communication, Cancer
Prevention and Control, 1999;3:25-30
Brown, Stewart & Ryan. Outcomes of
Patient-Provider Interaction, Handbook of
Health Communication, Clinics in
Geriatric Med, 2000;16:25-36
PHYSICIAN OUTCOMES

Malpractice claims

Time

Physician satisfaction
Malpractice Claims

8 studies

RCTs and other designs
Malpractice Claims

Important differences
eg. Hickson et al., 1994
% Poor
Patient-Physician
Communication
No claims
All others
High frequency
8.2
17.7
27.6
High pay
24.7
Malpractice Claims
1) time: feeling rushed, short visit
2) explanation: inadequate/minimal information, fewer
orienting statements
3) connection: feeling ignored, no acknowledgement of
patient statements, no reflection of affect, no eye
contact, no friendly physical contact, no humour
4) facilitation: not understanding patient and family
perspectives, no eliciting patients’ opinions and cues
5) support: devaluing patient and family views, harsh
critical tones
Physician Outcomes

Malpractice claims

Time

Physician satisfaction
Time

19 studies

RCTs and other designs
Time

6 studies found no differences in time
but communication was increased in
terms of

exploration of psychosocial issues

collaboration with patients

exploring patients’ ideas and
concerns

patient volume of communication

patient-centred communication
Time

10 studies found differences in visits of
longer duration in terms of

more counselling

more prevention (e.g. vaccination)

larger proportion of patient needs
recognised

higher patient satisfaction

more follow-up

better guideline implementation
Time

family practice average 10 minutes

primary care internal medicine average 21
minutes
Time
In the situation of limited time, patient
satisfaction is related to:
• a brief period of time to "chat" about
non medical topics
• providing patients with feedback on
clinical findings
Physician Outcomes

Malpractice claims

Time

Physician satisfaction
Physician Satisfaction

one study

cohort study
Physician Satisfaction

Important difference in physician
satisfaction when the
communication was rated as
participatory
(Roter et al., 1997)
PHYSICIAN OUTCOMES

Malpractice claims

Time

Physician satisfaction
PATIENT OUTCOMES

Satisfaction

Adherence

Health
PATIENT OUTCOMES

Satisfaction

Adherence

Health
Satisfaction

many studies, none were RCTs

3 key review papers
Satisfaction

Consistent and important effects
of communication on patient
satisfaction
Satisfaction
1) Warmth/caring
2) Medical competence
3) Balanced communication of both
psychosocial and biomedical
concerns
4) Continuity of relationship
5) Expression of patient expectations
PATIENT OUTCOMES

Satisfaction

Adherence

Health
Adherence

16 studies and review papers

No RCTs
Adherence

Communication is the most important
determinant of patient adherence
e.g. Stewart, 1984
Physician
Patient-Centered
% Adherence
Low
55.4
High
73.1
Adherence
1)
information exchange and
patient education
2)
finding common ground
regarding expectations
3)
active role for the patient
4)
positive affect, empathy and
encouragement
PATIENT OUTCOMES

Satisfaction

Adherence

Health
Health Outcomes

23 studies

12 RCTs: 11 demonstrated
significant effects
Health Outcomes
(Kaplan & Greenfield, 1989)

Important effects
e.g. BP (diastolic)
Experimental
Control
Pre
95
93
Post
83
91
Health Outcomes

Important effects
e.g. Pain (Egbert et al., 1964)
Experimental Control
Severity of Pain
1
1.2
1.7
Health Outcomes Affected






Patient anxiety
Psychologic Distress
Symptom Resolution
Functional Status
Self-reported Health Status
Physiologic status e.g.
HA1
BP
“If a new drug was shown to
be as effective in rigorous
studies as patient-physician
communication is, the
industry would aggressively
market that drug!”
Health Outcomes
Physician during history taking

asks many questions about
patient’s ideas and expectations

asks about patient’s feelings

shows support
Health Outcomes
Physician during discussion of the
management plan

encourages patient to ask
questions and get information

provides information packages

provides emotional support

willing to share decisions

physician and patient agree on the
nature of the problem and the follow-up
PATIENT OUTCOMES

Satisfaction

Adherence

Health
Does Patient-Physician
Communication Matter?
Yes - to the physician
Yes - to the patient
Evidence-based guidelines for
patient-physician communication

Clear information provided to
the patient

Asks about patient ideas,
expectations and feelings

Mutually agreed upon goals

An active role for the patient

Positive affect, empathy, and
support from the physician
Hour 1 - Small Group Discussion
a) quality of the evidence presented
b) effective communication components
Hour 2
Mini-lecture on how to evaluate
cohort designs
5 min
Lecture on A Study of the
"Impact of Patient-Centred
Communication on Patient
Outcomes in Family Practice”
- a cohort study
20 min
Small group discussion
- design your own cohort study
20 min
Break
15 min
Categories of Research Design

Observational or experimental

In an observational study the
researcher collects information but
does not influence events
(Level II - 2)

By contrast, in an experimental
study the researcher deliberately
influences events (Level I)
Observational studies

cohort design

case-control study

cross-sectional study
Cohort study

Selection of subjects

Loss to follow-up

Other problems

Long-term studies may suffer
from problems associated with
changes and over time

Surveillance bias
THE IMPACT OF
PATIENT-CENTRED
CARE ON PATIENT
OUTCOMES IN FAMILY
PRACTICE
Grant #04106
Health Services Research, Ministry of Health of Ontario
Investigators:
Moira Stewart, Ph.D.
Judith Belle Brown, Ph.D.
Allan Donner, Ph.D.
Ian R. McWhinney, M.D.
Julian Oates, M.D.
Wayne Weston, M.D.
HYPOTHESIS
That adult patients whose first visit in an episode of illness is
patient-centred will more frequently demonstrate recovery
from the discomfort of the symptom after two months (and
recovery from the concern about the symptom) and will
experience less subsequent medical care (i.e. fewer patientinitiated visits, fewer tests and referrals) in the two months of
study.
PATIENT-CENTRED MEASURES

Measures of Patient-centred
communication based on analysis of
audiotapes (continuous)

Patient perception of patientcentredness total score (continuous)
PATIENT-CENTRED MEASURES

Patient perception, subscore that the
illness experience had been explored
(continuous)

Patient perception, subscore that the
patient and doctor found common
ground (continuous)
PATIENT HEALTH OUTCOMES

Patients’ level of discomfort, postencounter and two months later
(continuous)

Patients’ level of concern, post-encounter
and two months later (continuous)
PATIENT HEALTH OUTCOMES
Medical Outcomes Study (MOS) variables
assessed two months after the study visit:

physical health (continuous)
mental health (continuous)
perceptions of health (continuous)
social health (continuous)
pain (continuous)

role function (dichotomous)




MEDICAL CARE OUTCOMES
Lab tests ordered during the two months
(dichotomous)
 Referrals during the two months
(dichotomous)
 Number of visits during the two months
after the study visit (continuous)

ANALYSIS
confounding variables assessed;
 confounding variables included;
 multiple regression for continuous
outcomes adjusting for practice using
PROC MIXED
 multiple logistic regression for
dichotomous outcomes adjusting for
practice using PROC LOGISTIC and
PROC IML

RESULTS
random sample of physicians
 52% refusal rate
 participants were:
– same year of graduation and location
of practice
– more likely to be CCFP
 n = 39

RESULTS
28% patient refusal rate
 participants were:
– same age as all eligibles
– more likely to be male than all
eligibles
 n = 315

DEMOGRAPHIC
CHARACTERISTICS
OF THE PATIENTS
54% female
 40% > 45 years of age
 60% married
 42% had some post-secondary education

Multiple Regression of Perception Scores
in Relation to Patients’ Discomfort
n = 297
Outcome - Patients’ Level of Discomfort
B Estimate
SE
p
Total Patient Perception Score
6.04
2.70
0.03
Baseline level of discomfort
0.84
0.037
0.0001
–Musculoskeletal
–Respiratory
–Other
–Digestive
2.42
6.56
2.42
6.18
3.39
3.25
3.24
4.07
0.48
0.04
0.46
0.13
Patients’ marital status
0.63
2.03
0.76
Independent Variables:
Patients’ main presenting problem
Mean Level of Discomfort by Quartiles of
the Total Patient Perception Score
Perceptions:
First quartile - perception that the
visit was patient-centred
Second quartile
Third quartile
Fourth quartile - perception that the
visit was not patient-centred
Mean Level
of Discomfort
x
42.5
45.0
45.2
48.8
Proportion Receiving Diagnostic Tests by
Quartiles of the Score on Patient Perception of
Patient Centredness
Perceptions:
Proportion
Receiving
Diagnostic Tests
First quartile - perception of finding common
ground
Second quartile
Third quartile
Fourth quartile - perception of not finding
common ground
14.6%
17.0%
19.5%
24.3%
Proportion Referred by Quartiles on
Patients’ Perception of Patient
Centredness
Perceptions:
First quartile
- perception of finding common ground
Second quartile
Third quartile
Fourth quartile
- perception of not finding common ground
Proportion
Referred
7.9%
4.3%
6.9%
16.2%
Measure of Patient-Centred Communication
OUTCOMES
Measure of patientcentred
communication
Patient perception
of patientcentredness
Patients’ level of discomfort
NS
S
Patients’ level of concern
NS
NS
SF-36
NS
S
Diagnostic tests
NS
S
Referrals
NS
S
Number of subsequent visits
NS
NS
Relationships Among Measure of Patient-Centred
Communication, Patient Perceptions of PatientCentredness, and Outcomes
Measure of PatientCentred Communication
N.S.
N.S.
Health and Medical
Care Outcomes
Significant p=.01
Significant p<.05
N.S.
Patient Perception that
doctor explored the illness
experience
Patient perception that
the doctor and patient
found common ground
Surprising Result
The communication score (what we
think is good communication) was not
related to outcomes.
BUT
The patients’ perception (what the
patients noticed) was related to
outcomes.
Conclusion
Only when physicians’ level of patientcentredness reaches a level that patients
notice, outcomes are affected.
Hour 2 - Small Group Discussion
Design your own cohort study on the topic
Hour 3
Mini-lecture on how to evaluate
clinical trial designs
5 min
Lecture on A Study “Innovative
Training to Improve Physician
Communication with Breast
Cancer Patients: Results of a
Randomized Controlled Trial”
15 min
Small group discussion
- design your own clinical trial study
15 min
Closing - a synthesis and summary of
take-home messages with examples of
exam questions
15 min
Intervention Studies - Validity

Was the assignment of patients to
treatment/placebo randomized?

true randomization

pseudo-randomization or deterministic
method of allocation

treatment/placebo allocation
Intervention Studies - Validity

Were all patients properly accounted for
and attributed at its conclusion?

completeness of follow-up

intention to treat analysis
Intervention Studies - Validity

Were patients, clinicians and study
personnel “blind” to intervention?

double blinding

methods used to ensure blindness

creative ways to “unblind” a study
Intervention Studies - Validity

Were groups similar at the start of the
trial?

carefully consider all important
baseline characteristics of both
groups

post-hoc analysis accounting for
differences in the groups
Intervention Studies - Validity

Aside from the experimental
interventions, were the groups
treated equally?


co-interventions
How large was the treatment effect?
Innovative Training to Improve
Physician Communication With
Breast Cancer Patients: Results of a
Randomized Controlled Trial
Principal Investigator: Moira Stewart, Ph.D.
Funded by:
Canadian Breast Cancer Research Initiative &
National Cancer Institute of Canada
Breast Cancer Survivors:
Louisette Smith
Brenda McKelvey-Donner
Research Expertise:
Judith Belle Brown, Ph.D.
Carol McWilliam, M.Sc.N., Ed.D.
Allan Donner, Ph.D.
Clinical Expertise:
Ron Holliday, M.D.
Ken Leslie, M.D.
Tim Whelan, M.D.
Alan Gavin, M.S.W.
Irene Cohen, M.D.
Marjorie Wood, MB,ChB
Praful Chandarana, MB,ChB
Education Expertise:
Wayne Weston, M.D.
Tom Freeman, M.D.
Susan McNair, M.D.
Consultants:
Ian R. McWhinney, M.D.
Jack Laidlaw, M.D.
Don Cowan, M.D.
Project Coordinator:
Joanne Galajda
Research Assistant:
Jo-Anne Aubut
Breast Cancer Advisory Group
Co-Chairs: Louisette Smith
Brenda McKelvey-Donner
Members: Sharron Bearfoot
Registered Nurse
Vice-Principal
R.N. at Coronary Care Unit
Barb Barton-McMillan
Social Worker
Anne Buchanan
Volunteer, Canadian Cancer Society
Katherine DeCaluwe
Hair Stylist
Barbara Garvin
Regional Manager, London Canadian
Cancer Society
Addie Gushue
Medical Claims Assessor for the
Ministry of Health
Sandy Krueger
Independent Business Woman
Margie McPhillips
Homemaker
Does intensive training improve
physician communication with
breast cancer patients?
Purpose of the Study
 To design an education program for
physicians based on real life
experiences of breast cancer
survivors, with the goal of improving
physicians’ communication
 To evaluate the program
3 Phases of Study
 1. Qualitative Study
 2. Pre-test education program
 3. Evaluate the education program
in a randomized controlled trial
Qualitative Study Findings
Relationship-building
Information Sharing
Creating the Experience of Control
Mastering the Whole Person
Overview of the 6 Hour CME
Patient-Physician Communication
•Introduction
•Physician’s Perspective
•Patient’s Perspective - Video
•Patient’s Perspective - Discussion With Breast Cancer
Survivors
•Video Demonstration
•Lunch
•Videotaped Standardized Patient Interviews (2)
•Video Feedback
•Evaluation of the Course
Phase 3 - Randomized Controlled Trial
Recruitment - 51 family physicians, surgeons, oncologists
Randomization
2 standardized
patient appointments
2 standardized
patient appointments
6 hour education
2 hour education
2 standardized
patient appointments
2 standardized
patient appointments
Oncologists & Surgeons Only
10 patient
questionnaires
10 patient
questionnaires
Outcome Measures
1.
Audiotaped standardized
patients in office setting
2.
Patient questionnaires
Measuring Communication
Building Relationships: 1. Eliciting & Validating Feelings
2. Eliciting & Validating Ideas
3. Offering Support
Sharing Information:
4. Opportunity to Ask Questions
5. Mutually Discussing the Problem & Management Plan
6. Clarification of Agreement
Creating the Experience of Control:
7. Eliciting & Validating Expectations
8. Eliciting & Validating Impact on Function
Mastering the Whole Person Experience:
9. Eliciting & Validating Issues of the Person
RESULTS
Physician Communication Scores
from Audiotaped Office Visits
(6 hour vs. 2 hour education)
p-value
Oncologists:
N.S.
Surgeons:
N.S.
Family Physicians:
Eliciting & Validating Issues of the Person
.02
(Validating Issues of the Person)
.05
Offering Support
.02
Eliciting and Validating Issues of
The Person*
Family Physicians
Pre
Post
72.0
82.8
60.1
58.7
6 hour education
(n = 8)
2 hour education
(n = 9)
Multiple regression:
* (p=.02) Statistically significant difference between the 6 hour education and the
2 hour education at post-intervention controlling for pre-intervention scores
Patient Outcomes
Differences in patient outcomes between oncologists
and surgeons who had 6 hr. education vs. those who
had 2 hr. education:
 Perceptions of patient-centred
communication
N.S.
 Satisfaction with Dr.’s informationgiving & interpersonal skills
.03
 Psychological distress
N.S.
 Feel better after visit with Dr.
.02
Patient Feeling Better
After Visit With Doctor*
6 hour education
(n = 51)
88.2%
2 hour education
(n = 51)
70.6%
Logistic Regression
* (p=.02) Statistically significant difference between the patients
whose doctor was in the intervention group vs. the control group
at post-intervention controlling for patient’s education and
number of medical conditions
Conclusions About Effectiveness of
the 6 Hour Education
 Family physicians improved
communication (audiotapes)
 Patients of oncologists & surgeons
had greater satisfaction & felt
better after visit
Hour 3 - Small Group Discussion
Design your own randomized controlled trial
on the topic.
Take-home messages
– there is an evidence-base for patient-centred
communication
– there are 5 key communication elements
revealed by the literature
– cohort studies have a role in explaining the
impact of patient-centred communication
– the cohort study revealed a pathway
– communication of practising physicians and
surgeons can be improved through CME
– the RCT revealed an effect on behaviour and
patient perceptions
Take-home messages
– Lessons from research need to take
account of quality of study design
(Level I-III and A to E)
– Lessons also need to take account of
clinical significance as well as statistical
significance
– There were 6 criteria presented to
evaluate the quality of cohort studies
– There were 6 criteria presented to
evaluate the quality of the RCT
Sample examination questions
1.
Which is a correct label for a Level of Study Design?
a) 100%
b) II - 1
c) 3 - 6
d) B
2.
Which is a correct label for a Recommendation for the
Evidence created by the Canadian Task Force on
Preventive Health Care?
a) 2
b) A
c) 80%
d) III
Sample examination questions
3.
Clinical significance means the chances the study
results could have occurred by chance alone.
a) True
b) False
4.
How many communication elements does the
literature reveal to affect outcomes?
a) too many to learn
b) 75
c) two
d) five
Sample examination questions
5.
Which is a key criterion to evaluate a cohort study?
a) measure the outcome first
b) measure the factor before the outcome
c) be sure the patients are “blind” to the intervention
d) be sure the subjects were selected over a long period of time
6.
What is the key criterion to evaluate an RCT?
a) follow patients over time
b) lose patients to follow-up
c) measure the outcome first
d) the assignment of participants is randomized