Helping Patients Make Better Medical Decisions:

1
Leeds Institute of Health Sciences FACULTY OF
MEDICINE HEALTH

• Helping Patients Make Better Medical Decisions
• The Evidence For Decision Aid Interventions.
• Dr Hilary Bekker, PhD, MSc, BSc.
• Chartered Health Psychologist
• Senior Lecturer in Behavioural Sciences

2
Overview.

• Evidence from decision aid and prenatal testing
  research.
• Why support patients to make medical decisions?
• An example from the prenatal diagnosis context.
• What type of decision support?
• Where should clinical practice go from here?
• Where should research go from here?

3
Key Reviews of Primary Evidence.

• Informed decision making an annotated
  bibliography and systematic review. HTA. 1999.
  Bekker et al. (547 trials)
• Decision aids for people facing health treatment
  or screening decision. Cochrane. 2003. OConnor
  et al. (34 trials)
• Patient Participation in the consultation
  process a structured review of intervention
  strategies. PEC. 2006. Haywood et al. (137
  trials)
• A systematic review of patient needs in making
  difficult and proxy decisions. Under review.
  2007. Jackson C et al. (173 surveys)

4
Why support patient decision making?

• Policy informed decisions, shared decisions,
  expert patient, patient choice, and patient
  participation.
• Treatment options risks of serious
  consequences.
• Human decision making limited capacity to make
  difficult decisions well subject to biases when
  processing information.
• Aversion humans prefer not, or lack skills, to
  make decisions well - effortful, time-consuming
  and emotionally unpleasant.
• Patient outcome ?decisional conflict and regret
  ?informed decisions ?realistic expectations
  ?decisional certainty ?decision.

5
Real-World ExampleTo have or not have prenatal
diagnosis.

• During pregnancy, screening test to know risk of
  baby having Downs syndrome (DS).
• If screen positive, offered a diagnostic test
  (amniocentesis or chorionic sampling) to know for
  certain if baby has DS.
• Option 1 Option 2
• Continue Pregnancy Have Test
• Risk Downs Risk Miscarriage
• (1 in 5 1 in 250) (1 in 50 1 in 200)

6
Human processing exampleBiased interpretation
of risk figures.

• 1 in 50 babies are born with an abnormality.
• Your risk of the baby having Downs is 1 in 150.
• Do you think your risk low or high?
• If your triple test risk is ? 1 in 250 you will
  screen positive and be offered a diagnostic test.
• Your risk of having a baby is 1 in 150.
• Do you think your risk is low or high?

7
Human processing exampleScreening out
alternatives.

• 44 consultations between women and professional.
  Noted whether mentioned topic or not.
• Alternatives No test 59
• Test 100
• Consequences Miscarriage 57
• Termination 71
• Downs Syndrome 52
• Risks Screening test 77
• Test Miscarriage 64

8
Decision aid example to have or not to have
prenatal diagnosis for DS.
Decision Tree Prompt
9
Decision aid example to have or not to have
prenatal diagnosis for DS.
Elicitation of Utilities Prompt
10
Decision aid example to have or not to have
prenatal diagnosis for DS.
Threshold Graph Prompt link utility and risk
information.
11
How women reason in a decision aided
consultation.

• When a decision aid is used during routine
  consultations
• Evaluate more information about options, pros
  and cons.
• More realistic evaluation of screening test risk
  figure.
• Used more thinking phrases (metacognition).
• More confident with the choice they made over
  time routine group women were less confident
  over time.
• Express more emotions.
• More realistic evaluations of consultation
  information.
• Slightly longer consultations (6 minutes)

12
What Type of Decision Support ?

• Depends on the goal of the intervention
• To inform and/or prepare and/or educate
• (good quality, complete information, include
  risks benefits)
• To participate as an informed decision maker
• ( explicit reference to decision, value
  elicitation, prompts)
• To participate as a shared decision maker
• ( clinician prompt, clinician training in
  negotiation)

Passive role
Active role
13
Implications for clinical practice.

• Essential for educating, preparing and decision
  making.
• Written information
• Readable.
• Comprehensive information about all options (e.g.
  treatment and no treatment options)
• Risk presented in a way that reduces bias
• Pros and cons described of each option described
  
• No value terms (e.g.small risk).

• Essential for decision making.
• Prompts and/or aids
• Make explicit which decision(s) to be made and
  what the decision options and consequences are
• Make explicit whose responsibility the decision
  is
• Elicit values about options and consequences
• Assist reasoning, e.g. prepare questions
  decisional balance table structuring information
  by attribute.

14
Implications for Research.

• Introduction
• Development aids correct aid for goal chronic
  versus acute choices proxy versus self etc.
• Evidence decision aid effectiveness compared
  with routine, simple and complex aids.
• Implementation integrating aids in routine
  practice professional and patient training.

• Maintenance.
• Update aids need to accommodate new evidence and
  new decision aid techniques.
• Training health professionals - undergraduate
  and postgraduate education.
• Long-term evidence effectiveness future
  choices different treatment decisions.

15
Concluding comments.

• Humans do not make difficult decisions well,
  employ heuristic rather than systematic systems
  to process information.
• Many interventions can help humans to make
  decisions well.
• Components of interventions will differ depending
  on the goal of the situation.
• Developing decision aids requires
  multidisciplinary and multi-user input (e.g.
  understanding how people make decisions,
  professional practice, and patient needs).
• Research required to develop reliable evidence on
  what type of interventions are effective in which
  contexts always should be good quality but not
  always complex interventions.