Measurement of Patient Preferences – GW4580M
Inhoudsopgave
Week 1 ................................................................................................................................................................................ 1
Lecture Theoretical foundation, course overview, and survey development ................................................................................. 1
Week 2 ................................................................................................................................................................................ 7
Literature .................................................................................................................................................................................. 7
Lecture experimental DCE designs ............................................................................................................................................. 7
Practical Design Generation tutorial ........................................................................................................................................ 17
Week 3 .............................................................................................................................................................................. 25
Literature ................................................................................................................................................................................ 25
Lecture Discrete choice modelling (MNL) ................................................................................................................................. 25
Practical C: Discrete choice modelling using Stata ................................................................................................................... 37
Week 4 .............................................................................................................................................................................. 51
Literature ................................................................................................................................................................................ 51
Lecture Discrete choice modelling (Latent Class) ...................................................................................................................... 51
Practical D: Interpretation of DCE model output .................................................................................................................... 65
Week 5 .............................................................................................................................................................................. 69
Literature ................................................................................................................................................................................ 69
Lecture Discrete choice modelling (Mixed Logit) & Choice-share predictions using DCE's ......................................................... 69
Practical E: MIXL model & choice predictions .......................................................................................................................... 79
Week 6 .............................................................................................................................................................................. 87
Lecture: Q&A ......................................................................................................................................................................... 87
Exam questions ................................................................................................................................................................. 89
Week 1
Lecture Theoretical foundation, course overview, and survey development
Learning goals
1. Describe the theoretical foundation(s) of discrete choice experiments
2. Recognize the importance of patient preference measurements in health care
1
, 3. Develop one's own DCE using state-of-the-art software and methods
4. Understand and be able to assess recent articles in the DCE literature
DCE applications in health and health care
• Measurement of patient preferences is a young field, but with increasing volume and recognition
• Key applications in health:
1. Improve patient-centered care e.g. ‘shared decision making’
2. Provide direction for development of future care
3. Extend traditional health technology assessments
4. Estimating utility weights within QALY framework
Patient preferences are very diverse: no injections, maximum effectiveness, minimal impact on daily life
Physician preferences: optimal medical treatment (with much less emphasis on daily impact)
Example 1: SDM
Some decisions are too complex and require decision support tools
• So-called ‘decision support tools’ convey (average or personalized) risk/benefit information
• But patient preference measurement itself is often left for unstructured and inefficient pre-surgical consultations
• Important role for explicit patient preference measurement:
• Provides physician with preference information
• Forces patients to carefully examine their preferences
Example 2: elective surgery
Non-surgical options
• Radiation therapy
• external beam, intra-operative, brachytherapy
• partial breast, intensity-modulated, proton therapy
• Chemotherapy
• Hormonal therapy
• Targeted therapy
• Surgical options
• Lumpectomy
• Mastectomy
• Different types
• With and without reconstruction
• Different types of reconstruction
Example 3. HTA for erectile disfunction treatment
1. Viagra
2. Vacuumpump
3. Injection
Example 4. Estimating EQ-5D QALY Tariffs
2
,Traditional preference elicitation method is time trade-off (TTO)
• Many disadvantages (see next slides)
• ‘Next generation’ preference elicitation method is based on discrete choice experiments
• DCEs are less complex, makes it possible to include many more elicitation tasks, which accommodates more realistic
statistical models (e.g. without constant proportional time preferences)
• DCEs also have a solid theoretic foundation. Let’s walk through several example preference elicitation tasks (i.e.
TTO, best-worst scaling, Thurstone scaling, DCE)
EQ-5D Attributes and levels
• Health is divided into 5 domains: • mobility • self-care • usual activities • pain & discomfort • anxiety & depression
• Each domain has five levels, ranging from: “no health problems” (1) to “extreme health problems” (5)
• Best possible health state = 1-1-1-1-1 • Worst possible health state = 5-5-5-5-5
• 3,125 health states in total.
Two example EQ-5D health states
TTO explained
• Principle underpinning TTO: the worse the health state, the more time in
full health (1-1-1-1-1) someone would be willing to sacrifice in order to
avoid it. Respondents are required to converge (i.e. iterate) towards
their point of indifference (x).
Suppose the respondent is indifferent between 10 years in the impaired
health state and 9 years in perfect health. The health state value is x/t
= 9/10 = 0.9.
Obtaining the QALY tariff
Either let respondents evaluate all 3,125 health states, or evaluate a subset
and use a regression analysis (e.g. OLS or Tobit regression) to interpolate preferences. The regression results are the EQ5D
QALY tariff.
Problems with TTO
1) Complex • cannot be conducted in unattended surveys ($$)
2) Interviewers can influence results • requires training and quality control system
3) Few choice tasks per respondents feasible • Low statistical power • Statistical models impose linear time preferences
(unrealistic)
4) Health states worse than death are not possible • Requires more difficult TTO based on additional assumptions • for
example, with 10 years lead time (which places some evaluations 10 years into the future)
Alternative methods
Origin of DCEs: Thurstone scaling
• DCEs originated in the 1970s
• Thurstone, however, already proposed his “Law of Comparitive Judgement” in the 1920s
• Thurstone (‘27) was the first to relate observed choice probabilities to utility of the options in paired comparisons
Assumptions:
1) choice probabilities reflect distances between options on a latent scale;
2) preferences are distributed (normally) around the modal preference
• Thurstone scaling: the first technique in psychology for measuring latent constructs like preference
3
, Only feasible for a limited number of objects
Thurstone used pairwise comparisons to derive scale values for any arbitrary
set of objects
• Valuing 7, 10, or 20 objects requires respectively 21, 45, 190 comparisons (= n*(n-1)/2)
• Number of items is thus restricted to what you can handle in a paired comparison study
• Not possible to extrapolate results to other objects
Identifying attributes and levels
• What makes you choose one beer over another?
Please take surveys now (10 minutes)
Thurstone scaling: 15 questions to value 6 health states
• DCE: 10 questions to value 243 (minimum requirement: one choice set per parameter in the utility function).
• BWS: (minimum requirement: depends on design)
Discussion
• Which was the easiest survey to complete?
• Which was the paired comparison for Thurstone scaling, which the BWS, and which the DCE?
• How helpful was the color coding & level overlap?
• Could Thurstone scaling use the same visual presentation as the DCE?
By DCE, you have attributes and levels and you can almost choose randomly. Many more choice options that you can configure =
important distinction between dce and Thurstone scaling.
Best-worst scaling (types I, II, and III)
4
Inhoudsopgave
Week 1 ................................................................................................................................................................................ 1
Lecture Theoretical foundation, course overview, and survey development ................................................................................. 1
Week 2 ................................................................................................................................................................................ 7
Literature .................................................................................................................................................................................. 7
Lecture experimental DCE designs ............................................................................................................................................. 7
Practical Design Generation tutorial ........................................................................................................................................ 17
Week 3 .............................................................................................................................................................................. 25
Literature ................................................................................................................................................................................ 25
Lecture Discrete choice modelling (MNL) ................................................................................................................................. 25
Practical C: Discrete choice modelling using Stata ................................................................................................................... 37
Week 4 .............................................................................................................................................................................. 51
Literature ................................................................................................................................................................................ 51
Lecture Discrete choice modelling (Latent Class) ...................................................................................................................... 51
Practical D: Interpretation of DCE model output .................................................................................................................... 65
Week 5 .............................................................................................................................................................................. 69
Literature ................................................................................................................................................................................ 69
Lecture Discrete choice modelling (Mixed Logit) & Choice-share predictions using DCE's ......................................................... 69
Practical E: MIXL model & choice predictions .......................................................................................................................... 79
Week 6 .............................................................................................................................................................................. 87
Lecture: Q&A ......................................................................................................................................................................... 87
Exam questions ................................................................................................................................................................. 89
Week 1
Lecture Theoretical foundation, course overview, and survey development
Learning goals
1. Describe the theoretical foundation(s) of discrete choice experiments
2. Recognize the importance of patient preference measurements in health care
1
, 3. Develop one's own DCE using state-of-the-art software and methods
4. Understand and be able to assess recent articles in the DCE literature
DCE applications in health and health care
• Measurement of patient preferences is a young field, but with increasing volume and recognition
• Key applications in health:
1. Improve patient-centered care e.g. ‘shared decision making’
2. Provide direction for development of future care
3. Extend traditional health technology assessments
4. Estimating utility weights within QALY framework
Patient preferences are very diverse: no injections, maximum effectiveness, minimal impact on daily life
Physician preferences: optimal medical treatment (with much less emphasis on daily impact)
Example 1: SDM
Some decisions are too complex and require decision support tools
• So-called ‘decision support tools’ convey (average or personalized) risk/benefit information
• But patient preference measurement itself is often left for unstructured and inefficient pre-surgical consultations
• Important role for explicit patient preference measurement:
• Provides physician with preference information
• Forces patients to carefully examine their preferences
Example 2: elective surgery
Non-surgical options
• Radiation therapy
• external beam, intra-operative, brachytherapy
• partial breast, intensity-modulated, proton therapy
• Chemotherapy
• Hormonal therapy
• Targeted therapy
• Surgical options
• Lumpectomy
• Mastectomy
• Different types
• With and without reconstruction
• Different types of reconstruction
Example 3. HTA for erectile disfunction treatment
1. Viagra
2. Vacuumpump
3. Injection
Example 4. Estimating EQ-5D QALY Tariffs
2
,Traditional preference elicitation method is time trade-off (TTO)
• Many disadvantages (see next slides)
• ‘Next generation’ preference elicitation method is based on discrete choice experiments
• DCEs are less complex, makes it possible to include many more elicitation tasks, which accommodates more realistic
statistical models (e.g. without constant proportional time preferences)
• DCEs also have a solid theoretic foundation. Let’s walk through several example preference elicitation tasks (i.e.
TTO, best-worst scaling, Thurstone scaling, DCE)
EQ-5D Attributes and levels
• Health is divided into 5 domains: • mobility • self-care • usual activities • pain & discomfort • anxiety & depression
• Each domain has five levels, ranging from: “no health problems” (1) to “extreme health problems” (5)
• Best possible health state = 1-1-1-1-1 • Worst possible health state = 5-5-5-5-5
• 3,125 health states in total.
Two example EQ-5D health states
TTO explained
• Principle underpinning TTO: the worse the health state, the more time in
full health (1-1-1-1-1) someone would be willing to sacrifice in order to
avoid it. Respondents are required to converge (i.e. iterate) towards
their point of indifference (x).
Suppose the respondent is indifferent between 10 years in the impaired
health state and 9 years in perfect health. The health state value is x/t
= 9/10 = 0.9.
Obtaining the QALY tariff
Either let respondents evaluate all 3,125 health states, or evaluate a subset
and use a regression analysis (e.g. OLS or Tobit regression) to interpolate preferences. The regression results are the EQ5D
QALY tariff.
Problems with TTO
1) Complex • cannot be conducted in unattended surveys ($$)
2) Interviewers can influence results • requires training and quality control system
3) Few choice tasks per respondents feasible • Low statistical power • Statistical models impose linear time preferences
(unrealistic)
4) Health states worse than death are not possible • Requires more difficult TTO based on additional assumptions • for
example, with 10 years lead time (which places some evaluations 10 years into the future)
Alternative methods
Origin of DCEs: Thurstone scaling
• DCEs originated in the 1970s
• Thurstone, however, already proposed his “Law of Comparitive Judgement” in the 1920s
• Thurstone (‘27) was the first to relate observed choice probabilities to utility of the options in paired comparisons
Assumptions:
1) choice probabilities reflect distances between options on a latent scale;
2) preferences are distributed (normally) around the modal preference
• Thurstone scaling: the first technique in psychology for measuring latent constructs like preference
3
, Only feasible for a limited number of objects
Thurstone used pairwise comparisons to derive scale values for any arbitrary
set of objects
• Valuing 7, 10, or 20 objects requires respectively 21, 45, 190 comparisons (= n*(n-1)/2)
• Number of items is thus restricted to what you can handle in a paired comparison study
• Not possible to extrapolate results to other objects
Identifying attributes and levels
• What makes you choose one beer over another?
Please take surveys now (10 minutes)
Thurstone scaling: 15 questions to value 6 health states
• DCE: 10 questions to value 243 (minimum requirement: one choice set per parameter in the utility function).
• BWS: (minimum requirement: depends on design)
Discussion
• Which was the easiest survey to complete?
• Which was the paired comparison for Thurstone scaling, which the BWS, and which the DCE?
• How helpful was the color coding & level overlap?
• Could Thurstone scaling use the same visual presentation as the DCE?
By DCE, you have attributes and levels and you can almost choose randomly. Many more choice options that you can configure =
important distinction between dce and Thurstone scaling.
Best-worst scaling (types I, II, and III)
4
