Use of stated preference methods to quantify the determinants of medication nonadherence Rachel A El

1
Use of stated preference methods to quantify the
determinants of medication non-adherence Rachel
A ElliottBPharm MRPharmS PhDClinical Senior
LecturerHarkness Fellow in Health Care Policy
2004-5
2
Outline of seminar

• Non-adherence deviance or reasoned
  decision-making
• Discrete choice experiments an overview of a
  stated preference method
• Use of DCE in non-adherence rationale and study
  outline

3
1 Non-adherence deviance or reasoned
decision-making

• Non-adherence stems from two different types of
  behavior unintentional, intentional
• Predictors of non-adherence
• Demographic sex, age, income, ethnicity,
  education, payer/reimbursement
• Disease characteristics
• Drug characteristics
• Patient-prescriber relationships
• Is non-adherence a homogeneous phenomenon?

4
1.1 Non-adherence disease-specific rates
DiMatteo MR. Medical Care 2004 42 200-9
5
1.2 Determinants of non-adherence

• padhere ß1perceptions effectiveness ß2side
  effects ß3convenience ß4stigma ß5current
  symptoms ß6future disease risk ß7cost
  ß8relationship with clinician . ßiai e
• Different drug/disease pairs ? different
  coefficients for each parameter
• What happens with gt1 illness and/or gt1 drug?
• Complex trade-offs under conditions of scarcity
• Adherence levels reflect utility maximization
  behavior..how are these revealed preferences
  constructed?

6
2 DCE an overview of a stated preference method

• Discrete Choice Experiments (DCEs)
• Are a tool to elicit qualitative preferences and
  to quantify the value of preferences
• A good or service can be described by its
  attributes or properties, attributes are assigned
  levels.
• Utilitycar ß0 ß1fuel consumption ß2engine
  size ß3brand ß4capacity . ßiai e
• Attributes of the intervention are developed from
  interview/literature review, are assigned levels.

7
2. DCE an overview of a stated preference
method (cont.)

• The attributes and levels are combined to produce
  hypothetical scenarios
• Design optimisation criteria used to pair
  scenarios
• Respondents are presented with paired
  hypothetical scenarios and asked to select a
  preferred option (consistency tests included)
• This method asks respondents to trade off
  attributes of the service
• Logistic regression techniques used to analyse
  responses

8
2.1 DCE for drug treatment for BPH Watson V et
al. J Urol 2004 172 2321-2325
9
2.1 DCE for drug treatment for BPH Watson V
et al. J Urol 2004 172 2321-2325

• Choose A/B ß0 ß1a1 ß2a2 ß3a3 . ßiai e
  u
• A/BBPH ß0 ß 1TIME ß2SIZE ß3RED ß4IMP
  ß5ABEJ ß6HEAD ß7DIZZ ß8COST ß9AUR
  ß10SURG
• Marginal rates of substitution (MRS)
• WTP to avoid impotence 120
• WTP to avoid dizziness 75
• Willing to wait 13 months for ? prostate size
• Willing to wait 37 months for ? impotence

10
3 Use of DCE in non-adherence Rationale and
study outline

• V ß0 ß1a1 ß2a2 ß3a3 . ßiai e u
• Where a1 to ai are perceptions of drug
  effectiveness, side effects, convenience, stigma,
  current disease symptoms, future disease risk,
  relationship with clinician, cost
• Different drug/disease pairs ? different
  coefficients (part utilities) for each attribute
• What are the attributes and levels that need to
  go into the model?

11
3.1 CRUM Study qualitative study of determinants
of non-adherence

• How do the preferences of people with multiple
  chronic illness affect their decision to be
  non-adherent to medications for some or all of
  their conditions?
• Semi-structured interview with 20 community
  dwelling seniors with gt3 chronic medications
• Themes Understanding of disease, risk, impact on
  QoL, drug effectiveness, side effects, ease of
  use, relationship with prescriber, payer, cost,
  Significant others, trade-offs

12
3.1 CRUM Study qualitative study of determinants
of non-adherence

• If there was one of your medicines you would
  really like to get rid of, for whatever reason,
  which one would it be?
• Why did you pick that one? Are there any other
  reasons?
• If there was one of your medicines you would
  really not want get rid of, for whatever reason,
  which one would it be?
• Why did you pick that one? Are there any other
  reasons?