PatientPhysician Agreement on the Content of CHD Prevention Discussions

1
Patient-Physician Agreement on the Content of CHD
Prevention Discussions

• Lindy Behrend, MPH
• Hossein Maymani, BS
• Megan Diehl, BS
• Ziya Gizlice, PhD
• Jianwen Cai, PhD
• Stacey L. Sheridan, MD, MPH
• The University of North Carolina at Chapel Hill
• Funding Source National Heart Lung and Blood
  Institute (1 K23 HL074375)

2
Background

• To optimize decision making and clinical care, a
  common perception of the content and outcomes of
  clinical discussions is desirable.
• Little is known about whether patients and
  physicians perceive the content and outcomes of
  clinical discussions similarly.

3
Objectives

• To examine the agreement between patients and
  physicians regarding the content and outcome of
  coronary heart disease (CHD) prevention
  discussions.
• To compare patient and physician reports of
  content and outcome to coded transcriptions of
  clinic visits.

4
Methods Study Overview
Audio recording of visit
Patient and physician surveys about visits
5
Study Setting

• One university internal medicine clinic
• 94 providers (18 attendings and 76 residents)
• 59 (63) providers agreed to participate in the
  larger study
• 19 providers (16 attendings and 3 residents) had
  patients enrolled in the study at the time of
  this analysis

6
Participants

• All patients (n115) actively participating in
  the larger trial at the time of this analysis.
• Patients included if
• Ages 40-79
• No prior history of cardiovascular disease,
  diabetes mellitus, or other serious medical
  condition
• At moderate (6-9) to high risk (gt10) of heart
  disease over 10 years based on a Framingham risk
  equation
• Patients excluded if
• Presenting for first visit
• No cholesterol checks within past 18 months
• Unable to speak or read English
• Systolic BP gt180 or total cholesterol gt300

7
Study Procedures

• Following one clinic visit, we surveyed both
  patients and their physicians about visit content
  and outcomes.
• We audio-recorded a sub-sample of visits
• transcribed them verbatim
• two independent readers code their content (kappa
  0.88-1.0)

8
Survey Content

• CHD prevention discussion (yes, no)
• Discussion content (mostly pros, pros cons,
  mostly cons)
• Involvement in Decision Making
• Patient expression of preferences (a lot, a
  little, not really)
• Who made the final decision (MD, shared, patient)
• Recommendations and Final Decisions
• Physician recommendations (take medicine, change
  lifestyle)
• Final decision (take medicine, change lifestyle)

9
Analysis

• For each outcome of interest, we calculated
• Percent agreement
• Simple or Weighted Cohens kappas
• Almost perfect agreement 0.80 to 1.00
• Substantial agreement 0.60 to 0.79
• Moderate agreement 0.40 to 0.59
• Fair agreement 0.20 to 0.39
• Poor agreement 0.00 to 0.19
• No agreement lt0.00 (e.g. worse than chance)
• Examined patterns of disagreement

Landis and Koch 1977. Biometrics 33159.
10
Results Patient Characteristics
11
Physician Characteristics
12
Patient and Physician Agreement on Presence of
CHD Discussion (n115)
13
Patient and Physician Agreement on Discussion
Content (n98)
Physician report
Patient report
Total percent agreement 62 Kappa .22 (-.02 -
.46)
14
Patient and Physician Agreement on Patient
Expression of Preferences (n98)
Physician report
Patient report
8
Total percent agreement 43 Kappa .20 (.04 -
.36)
15
Patient and Physician Agreement on Who Made
Final Decision (n98)
Physician report
Patient report
Total percent agreement 44 Kappa .04 (-.10 -
.19)
16
Patient and Physician Agreement on
Recommendations and Final Decisions (n98)
17
Sub-sample Analysis
18
Verification of Discussion Presence (Sub-sample)
19
Verification of Discussion Content (Sub-sample)
20
Verification of Patient Involvement (Sub-sample)
21
Verification of Physician Recommendations and
Final Decisions (Sub-sample)
22
Conclusions

• Our analysis of patient-physician surveys
  indicated
• Fair to moderate agreement on physician
  recommendations and final decisions to take
  medicine/change lifestyle
• Poor to fair agreement on discussion content,
  patient expression of preferences, and who made
  final decision
• Coded transcriptions agreed with surveys on most
  outcomes
• Patient and physician surveys vs. transcriptions
  did not agree on discussion content or patient
  involvement

23
Limitations

• Small sample size overall limits precision of
  estimates
• Results from coded transcriptions are hypothesis
  generating only
• Potential for recall bias as physicians may not
  have completed surveys immediately following
  clinic visits
• Results may not generalize to more diverse
  populations, including those with less education
  or less desire for shared decision making

24
Implications

• Disagreements about content and participation in
  clinical discussions may be common.
• Audio recorded measures should be considered as
  an alternative to patient and physician reports
  of clinical discussions.
• Further study is needed
• in larger and more diverse populations
• to determine the impact on clinical outcomes
• to explore how best to facilitate clearer
  communication among patients and physicians

25
Thank you!
26
(No Transcript)
27
Where are the Disagreements Who Made Final
Decision
Patients perceived themselves as less involved in
DM than physicians perceived
Physician view of who made decision
Patient view of who made decision
Physicians and patients agreed on who made
decision
Patients perceived themselves as more involved in
DM than physicians perceived
28
(No Transcript)
29
Coding of Visit Transcripts

• CHD prevention discussion
• Specific statement about lowering chances of CHD
• Discussion content
• Counted of pros and cons
• Categorized as mostly pros, pros cons, mostly
  cons

30
Coding of Transcripts

• Patient expression of preferences
• Counted preferences expressed
• Categorized as a lot (3 or more), a little (1-2),
  not really (0)
• Who made the final decision
• Counted whether prevention options were
  discussed, the MD and patients expressed
  preferences, decision was made
• Combined into 5 categories (MD alone, shared,
  patient alone)
• Recommendations and Final Decisions
• Specific statement of recommendation or plan to
  take medicine or make lifestyle changes

31
Transcription definition of decision-making
Element may or may not be present a Involves
some, but not all, elements b Must involve all
elements
32
Transcription coding of decision-making

• What characterized the discussion about lowering
  the patients chances
• of heart disease?
• The doctor made the decisions using all that's
  known about the ways to lower the chances of
  heart disease.
• The physician assumed total control of decision
  making, involving some form of an authoritarian
  statement. The physician did not offer treatment
  options or check the patient for his/her
  understanding, values, or agreement.
• The doctor made the decisions but strongly
  considered the patients opinion.
• The physician assumed control of decision making,
  but asked for the patients opinions and values,
  or checked for the patients knowledge and
  understanding of treatment options.

33
Transcription coding of decision-making

• What characterized the discussion about lowering
  the patients chances
• of heart disease?
• The doctor and the patient made the decisions
  together on an equal basis.
• Shared decision making is a negotiated event
  that involves both discussion and choice. To be
  considered a shared decision, the interaction
  must include all of the following
• discussion of treatment options (including pros
  and cons) or physician check that patient has
  adequate knowledge/no remaining questions about
  the facts
• opportunity for the patient to express values,
  concerns, and/or preferences, or physician check
  for these things
• opportunity for physician to make
  recommendation/express values
• making or deferring a decision

34
Transcription coding of decision-making

• What characterized the discussion about lowering
  the patients chances
• of heart disease?
• The patient made the decisions but strongly
  considered the doctor's opinion.
• The patient assumed control of decision making
  (I want to do this), but asked for the
  physicians input/opinion or physician expressed
  opinion regarding treatment choice.
• The patient made the decisions using all he/she
  knows or has learned about how to lower the
  chances of heart disease.
• The patient made a treatment decision without
  discussion of the physicians opinion. The
  patient sought no treatment information from the
  physician, did not ask the physician to help
  clarify values, and did not ask for agreement.