Transcript The burden of upper gastrointestinal endoscopy in patients

Patients’ preferences for preventive
osteoporosis drug treatment
EW de Bekker-Grob
ML Essink-Bot
WJ Meerding
HAP Pols
BW Koes
EW Steyerberg
Dept. Public Health, Dept. Internal Medicine, and Dept. General Practice,
Erasmus MC Rotterdam, the Netherlands
[email protected]
BACKGROUND
Osteoporosis
 Risk factor for fractures (low bone mass)
 75.000 osteoporotic fractures annually in NL
 15.000 hip fractures
 60.000 other fractures
 Age risk factor
 Preventive medication (bisphosphonates)
Study
 Active case finding important to identify patients who benefit from
preventive drug treatment.
 Are subjects willing to take preventive drug treatment?
AIM
1. To elicit relative weight that patients place on various aspects of
preventive osteoporosis drug treatment
2. To investigate whether high risk patients had other preferences
than low risk patients.
METHODS I
Respondents
Women aged over 60 years
Identified by case finding in 34 GP-practices in NL
Overrepresentation high risk (10-yrs risk hip # > 6%)
DCE
Respondents choose between different options described by drug
treatment attributes at different levels
METHODS II
Background information DCE
 Survey method
 Good and services can be described by their characteristics
(attributes)
 Used to estimate:
 Whether attribute is important
 Relative importance of attributes
 Trade-off between attributes
 Willingness to pay (monetary measure of benefit)
ATTRIBUTES and ATTRIBUTE LEVELS
Attributes
Way of drug administration
Effectiveness (% risk reduction
of hip fracture)
Side effects (nausea)
Total treatment duration (yrs)
Total out of pocket costs (€)
Attribute levels
Tablet once a month,
Tablet once a week,
Injection every four months,
Injection once a month
5, 10, 25, 50
Yes, No
1, 2, 5, 10
0, 120, 240, 720
CHOICE SETS
 Fractional factorial design (main effects only design)
 16 drug profiles
 Folder-over technique (01, 12, etc) for minimal overlap
Way of drug
administration
10-years risk reduction
of a hip fracture
Nausea (during two
hours after use)
Total treatment duration
Total cost to you
Which treatment
do you prefer?
Treatment A
Treatment B
No treatment
Tablet
once a week
Injection by GP
every 4 months
Not applicable
10%
25%
0%
Yes
No
No
2 years
€0
5 years
€ 120
0 years
€0
0 Treatment A
0 Treatment B
0 No treatment
ANALYSES
 Patient preferences  conditional logit regression
V = β0 + β1TABLETweekly + β2INJECTIONfourmonths +
β3INJECTIONmonthly + β4EFFTIVENSS + β5NAUSEA + β6TIME +
β7COST
 Time and monetary trade-offs  ratios of coefficients
 High vs low risk  conditional logit regression with interaction
RESULTS: RESPONDENTS
Group
Age
60-64
64-69
70-74
75-79
80 and older
Household
single
with partner
Eductional level
Low
Intermediate
High
a
All
patients
(%)
120
LR
patients
(%)
60
HR
patients
(%)
60
a
P-value
<0.001 *
28
18
20
28
26
21
11
17
10
1
7
7
3
28
25
<0.001 *
53
67
17
43
36
24
0.215
63
46
11
31
26
3
32
20
8
significant difference between low and high fracture risk patient groups
* significant at the 5% level
RESULTS: CONDITIONAL LOGIT REGRESSION
Attribute
Constant (drug treatment)
Drug administration (base level
tablet once a month):
tablet once a week
injection every four months
injection once a month
Effectiveness (10% risk reduction)
Side effect nausea
Treatment duration (1 year)
Cost (€100)
Beta
coefficient
1.23
<0.001*
0.81
1.66
-0.31
-0.21
-0.44
0.28
-1.10
-0.04
-0.15
<0.001*
0.027*
<0.001*
<0.001*
<0.001*
<0.001*
<0.001*
-0.45
-0.41
-0.64
0.23
-1.30
-0.06
-0.18
-0.17
-0.02
-0.25
0.34
-0.89
-0.02
-0.11
p Value
95% CI
* significant at the 5% level
Number of observations 5,589 (117 respondents x 16 choices x 3 options per choice, minus 27
2
missing values), Pseudo R = 0.1847, log pseudolikelihood = -1668.7
RESULTS: TRADE-OFFS
Attribute
Constant (no drug treatment)
Drug administration (base level
tablet once a month):
tablet once a week
injection every four months
injection once a month
Effectiveness
Side effect nausea
Treatment duration
Cost
Willingness to
adhere to the
drug treatment
longer (years)
32.7
8.2
5.7
11.7
7.5
29.0
3.9
WTP for
total
treatment
(€)
847
212
147
304
195
752
26
RESULTS: EXAMPLE
Bisphosphonate
 Weekly tablet
 Nausea
 Duration 5 years
Women prefer this drug treatment above no drug treatment,
if WTP is positive, thus
0 < €constant + €weekly tablet + €nausea + €time + €effectiveness
WTP = 847 – 212 – 752 – 5*26 + 20*X
600
If risk reduction is larger than
12%, than out-of-pocket
payment becomes acceptable
WTP (€)
400
200
0
0
10
20
Risk Reduction (%)
-200
30
40
LOW VS HIGH FRACTURE RISK PATIENTS
Attribute
Constant (drug treatment)
Drug administration (base level
tablet once a month):
tablet once a week
injection every four months
injection once a month
Effectiveness (10% risk reduction)
Side effect nausea
Treatment duration (1 year)
Cost (€100)
*significant at the 5% level
Coefficient of
low risk
patients
1.18
Coefficient of
high risk
patients
1.32
-0.36
-0.13
-0.45
0.23
-1.05
-0.03
-0.20
-0.26
-0.32
-0.45
0.34
-1.16
-0.05
-0.10
p-value
0.75
0.46
0.32
0.97
0.05 *
0.58
0.58
0.44
Number of observations 5,589 (117 patients (i.e. 58 low-risk + 59 high-risk patients) x 16 choices x 3
options per choice, minus 27 missing values), Pseudo R2 = 0.190, log pseudolikelihood = -1659
CONCLUSIONS
1. Women exhibited a very positive attitude towards preventive
osteoporosis drug treatment
2. Important message for policy decision-making on introduction of
active osteoporosis case finding at large scale (in addition to
cost-effectiveness considerations)
3. This study demonstrates feasibility of DCE in older patients
We would like to thank the Netherlands Organization for Health Research
and Development (ZonMw) for funding the research.