1 EQ-5D, HUI and SF-36 Of the shelf instruments….

2 Direct valuation

3 …or use validated questionnaires MOBILITY  I have no problems in walking about  I have some problems in walking about  I am confined to bed SELF-CARE  I have no problems with self-care  I have some problems washing or dressing myself  I am unable to wash or dress myself USUAL ACTIVITIES (e.g. work, study, housework family or leisure activities)  I have no problems with performing my usual activities  I have some problems with performing my usual activities  I am unable to perform my usual activities PAIN/DISCOMFORT  I have no pain or discomfort  I have moderate pain or discomfort  I have extreme pain or discomfort ANXIETY/DEPRESSION  I am not anxious or depressed  I am moderately anxious or depressed  I am extremely anxious or depressed

4 Validated questionnaires

The Rosser & Kind Index 5

The Rosser & Kind index  One of the oldest valuation  1978: Magnitude estimation  Magnitude estimation  PTO / VAS  N = 70 Doctors, nurses, patients and general public  1982: Transformation to “utilities”  On a 0.00 to 1.00 scale  Could be used for QALYs 6

1985: High impact article 7

8

 Survey at the celebration of 25 years of health economics in the UK (HESG): chosen most influential article on health economics 9

Criticism on the Matrix  Sensitivity  only 30 health states  The unclear meaning of “distress”  The involvement of medical personnel  No clear way how to classify the patients  into the matrix  Only British values  The compression of states in the high values 10

Value compression 11

New initiatives  Higher sensitivity (more then 30 states)  More and better defined dimensions  Other valuation techniques  Standard Gamble, Time Trade-Off, Visual Analogue Scale  Values of the general public  A questionnaire…  to allow patients to ‘self classify’ themselves  An international standard  to allow international comparisons  That is at that time “Europe” 12

13 Validated questionnaires

14 No longer value all states  Impossible to value all health states  If one uses more than 30 health states  Estimated the value of the other health states with statistical techniques  Statistically inferred strategies Regression techniques EuroQol, Quality of Well-Being Scale (QWB)  Explicitly decomposed methods Multi Attribute Utility Theory (MAUT) Health Utility Index (HUI)

15 Statistically inferred strategies  Value a sample of states empirically  Extrapolation  Statistical methods, like linear regression  = 1.00  =.70  = ?

16 Statistically inferred strategies  EuroQol  EQ-5D: 5 dimensions of health  245 health states  Quality of Well-Being scale (QWB)  4 dimensions of health  2200 health states plus 22 additional symptoms  SF-36  SF-6D: 6 dimensions of health  health states

17 Explicitly Decomposed Methods  Value dimensions separately  Between the dimensions  What is the relative value of: Mobility…... 20% Mood…….. 15% Self care.… 24%.  Value the levels  Within the dimensions  What is the relative value of Some problems with walking…… 80% Much problems with walking…... 50% Unable to walk…………………….10%  = 1 - (0.20 x ( )) = 0.96

18 Explicitly Decomposed Methods  Combine values of dimensions and levels with specific assumptions  Multi Attribute Utility Theory (MAUT) Mutual utility independence Structural independence

19 Explicitly Decomposed Methods  Health Utilities Index (Mark 2 & 3)  Torrance at McMaster  8 dimensions  Mark 2: health states  Mark 3: health states  The 15-D  Sintonen H.  15 dimensions  3,052,000,000 health states (3 billion)

20 Exercise  EuroQol EQ-5D (3 level)  SF-6D

21 More health states, higher sensitivity ? (1)  EuroQol criticised for low sensitivity  Low number of dimensions Development of EQ-5D plus cognitive dimension  Low number of levels (3) Gab between best and in-between level  Now development of 5 Level  EQ-5D-5L  No consensus in EuroQoL that more levels is always better…

EQ-5D-3L versus EQ-5D-5l

23 More health states, higher sensitivity ? (2)  Little published evidence  Sensitivity EQ-5D < SF-36 Compared as profile, not as utility measure  Sensitivity EQ-5D  HUI  Sensitivity  the number of health states  How well maps the classification system the illness?  How valid is the modelling?  How valid is the valuation?

24 5Lmore sensitive than 3L

25 More health states, more assumptions  General public values at the most 50 states  The ratios empirical (50) versus extrapolated  Rosser & Kind1:1  EuroQol EQ-5D-3L1:5  QWB1:44  EQ-5D-5L1:63  SF-361:180  HUI (Mark III)1:19,400  15D1:610,000,000  What is the critical ratio for a valid validation?

26 SF-36 as utility instrument  Transformed into SF6D  SG  N = 610  Inconsistencies in model  health states  regression technique stressed to the edge  Floor effect in SF6D

27 Collapsing levels SF-6D  Many levels are taken together  If PF=2decrement:  If PF=3decrement:  If RL=2decrement:  If RL=3 *decrement:  If RL=4 *decrement:

28 SF-6D loses a lot of levels  Levels in system and actual levels  PF6 5  RL 42  SF 55  PN 65  MH 54  VI 53  Levels in system:  6x4x5x6x5x5  Actual levels: 480  5x2x5x5x4x3

Conflicting evidence sensitivity SF-36 Liver transplantation, Longworth et al., 2001

30 EQ-5D-3L  Strong punts  Very sensitive in the low  Measures subjective burden (inside the skin)  Low administrative burden  Many translations  Cheap  Most used QALY questionnaire  Most international validations  Weak points  Only there levels per dimensions  Insensitive in the high regions

31 EQ-5D-5L  Strong punts  Very sensitive in the low  Measures subjective burden (inside the skin)  Low administrative burden  Many translations  Cheap  Weak points  No scorings algorithm yet  Might still not be as sensitive in the high regions

32 HUI  Strong punts  Sensitive  Measures objective burden (outside the skin)  Well developed proxy versions  Well developed child versions  Weak points  Expensive

33 SF-6D  Strong punts  Probably sensitive in the high regions  Often already include in trials (SF-36)  Cheap ….  Many translations  Weak points  Insensitive in the low regions  Only one validation study  Changed Standard Gamble Upwards shift of values

EQ-5D-Y  No scoring algorithm

35 Conclusions  More states  better sensitivity  The three leading questionnaires  have different strong and weak points