Minimally Important Difference

MID

Miyazaki.kikuko.4s@kyoto-u.ac.jp

2014 6 21

PRO

1. Minimally Important Difference MID

QOL

2. MID

3. MID

4. MID COSMIN

5. EORTC QLQ-C15-PAL MID

2014/6/21 8 CSP-HOR 2

Donabedian

structure

process

outcome

2014/2/22 QOL/PRO 3

QI

(Donabedian, JAMA, 1988)

Patient-

reported-

outcome

(PRO)

• QOL

QOL

• QOL

2014/2/22 QOL/PRO 4

QOL

• WHO

1948

•

QOL

• QOL

2014/2/22 QOL/PRO 5

Patient-reported outcome(s): PRO(s)

European Medicines Agency EMA (2005)

U. S. Food and Drug Administration (FDA) (2009 )

PRO

•

2014/2/22 QOL/PRO 6

QOL

1.

2.

QOL

3. QOL

4. QOL

5. QOL

2014/2/22 QOL/PRO 7

1. Minimally Important Difference MID

2014/6/21 8 CSP-HOR 8

VAS well-being

10 VAS 0.5

0.5

0.5

Ρ<0.05

2014/6/21 8 CSP-HOR

(Jaeschke et al, Control Clin Trials, 1989) 9

QOL MID

QOL/PRO 10

QOL

PROQOL QOL

From Measurement to Understanding

QOL

MID

2014/6/21 8 CSP-HOR

MID Minimally Important Difference

QOL

MID

The smallest difference in score in the domain of interest

2014/6/21 8 CSP-HOR QOL/PRO 11

(Jaeschke et al., Control Clin Trials,1989, )

• QOL

• QOL 5

5

• 5

• 5

•

2014/6/21 8 CSP-HOR 12

2014/6/21 8 CSP-HOR 13

2 MID

2014/6/21 8 CSP-HOR QOL/PRO 14

QOL

Distribution-based methods

- QOL

-

Anchor-based methods

-

2014/6/21 8 CSP-HOR QOL/PRO 15

Patient subjective view

PS,

Clinical parameter

MID

Distribution-based method Distribution-based method

Anchor-based method

Patient subjective view

PS,

Clinical parameter

Anchor-based method

Clinician perspective

MMMMMMMMMMMMMMIIIIIDDDDDDDDDDDDDMID

MID

QOL

2014/6/21 8 CSP-HOR 16

• Distribution-based method

• Anchor-based method

2014/6/21 8 CSP-HOR

17

3. MID

2014/6/21 8 CSP-HOR QOL/PRO 18

2014/6/21 8 CSP-HOR 19

Osoba et al. 1998

246 111

EORTC QLQ-C30 Global rating scale

- 5-10

- 10-20

- 20

2014/6/21 8 CSP-HOR 20

Norman et al. 2003

Systematic review

38 62 effect size

MID 0.5 SD

2014/6/21 8 CSP-HOR 21

Wyrwich et al. 2007

- COPD 2988 610

- CRQ Chronic Respiratory Disease Questionnaire SF36

-

- MID

4. MID COSMIN

2014/6/21 8 CSP-HOR QOL/PRO 22

COSMIN: COnsensus-based Standards for

the selection of health Measurement INstruments

HR-PRO

(http://www.cosmin.nl/)

2014/6/21 8 CSP-HOR 23

COSMIN • 2006 -2007

91 43

•

•

• [Classical Test Theory: CTT]

[Item Response Theory: IRT]

• 4

4 18

2014/6/21 8 CSP-HOR 24

Box J. Interpretability

8. Was the minimal important change (MIC)

or the minimal important difference (MID)

determined?

2014/6/21 8 CSP-HOR 25

5 EORTC QLQ-C15-PAL MID

2014/6/21 8 CSP-HOR QOL/PRO 26

Background Palliative Medicine

• Palliative care is an approach that improves the quality of life (QOL) of patients and their families facing the problem associated with life-threatening illness.

QOL (WHO, 2002, National Cancer Control Programmes

Policies and managerial guideline 2nd edition)

• Outcomes of palliative care can be obtained by measuring patient QOL.

QOL

2014/6/21 27 8 CSP-HOR

Purpose

To estimate MID for EORTC QLQ-C15-PAL scores for the assessment of QOL of palliative care patients.

QOL EORTC QLQ-C15-PAL(European Organisation for Research and Treatment of Cancer Quality of Life Core 15 Palliative) MID

2014/6/21 28 8 CSP-HOR

29

First survey 1 Second survey 2

2w

EORTC QLQ-C15-PAL

Clinical information*

EORTC QLQ-C15-PAL

GRS

Clinical information*

“What do patients themselves think of the

difference between the first and second time

they participated in the QOL survey?”

QOL

Measurement method

2014/6/21 8 CSP-HOR

EORTC QLQ-C15-PAL

• EORTC QLQ-C15-PAL was developed for palliative care patients by taking

the 30-item QLQ-C30 as a base and reducing the number of questions to

15 items.

EORTC QLQ-C15-PAL QLQ-C30

30 15 15

(Groenvold M, et al. Eur J Cancer, 2006)

2014/6/21 8 CSP-HOR 30

Results Flowchart for Patient Selection

Request participation

n=257

Number of patients enrolled in

the 1st QOL survey

1 QOL

n=240

2nd QOL survey and

GRS survey completed

MID analysis subjects 2

n=181

Refusal n=12

Worsened condition n=3

Transfer n=2

Death n=20

Worsened condition n=25 No follow-up due to transfer

n=11

Discontinuation by request

n=2

Unknown n=1

2014/6/21 31 8 CSP-HOR

2nd survey-

response(n=181)

2nd survey-no-

response(n=59)

χ2 test

p-value

Number of

patients

or average

or SD Number of

patients

or average

or SD

Sex Male 92 50.8% 38 64.5% 0.073

Female 89 49.2% 21 35.6%

Age 66.8 12.5 67.3 10.4 0.646

Underlying Lung cancer 31 17.1% 11 18.6% 0.535

disease Colon cancer 24 13.3% 6 10.2%

Brest cancer 21 11.6% 5 8.5%

Prostate cancer 14 7.7% 4 6.8%

Stomach cancer 12 6.6% 5 8.5%

Esophageal cancer 10 5.5% 2 3.4%

Uterine cancer 10 5.6% 3 5.1%

Pancreatic cancer 9 5.5% 6 10.2%

Renal cancer 8 4.4% 0 0.0%

Liver cancer 7 3.9% 7 11.9%

Others* 35 19.3% 10 16.9% 32

*Gallbladder cancer, peritoneal cancer, oral cancer, pharyngeal/laryngeal cancer, bladder cancer, Ewing’s sarcoma, etc.

Characteristics of the patients 2 2

2014/6/21 8 CSP-HOR

2nd survey-

response(n=181)

2nd survey-no-

response(n=59)

Number of

patients

Number of

patients

χ2 test

p-value

ECOG PS 0 10 5.5% 2 3.4% 0.014

(1st time) 1 30 16.6% 5 8.5%

2 61 33.7% 12 20.3%

3 67 37.0% 29 49.2%

4 13 7.2% 11 18.6%

ECOG PS 0 10 5.5%

(2nd time) 1 30 16.6%

2 60 33.1%

3 64 35.4%

4 17 9.4%

2014/6/21 33

ECOG PS of patients based on participation in

the second survey 2 2

8 CSP-HOR

Discussion

Examination of MIDs for each domain

• EORTC QLQ-C15-PAL

•

• QOL

2014/6/21 34 8 CSP-HOR

Direction toward improving or worsening

• Anchor-based MID

(Cella, et al. Qual Life Res, 2002)

• QOL MID

Limitation

QOL

2014/6/21 35 8 CSP-HOR

Conclusion

• MID 11.5

• MID 4.5

• QOL

MID

2014/6/21 36 8 CSP-HOR

1. Minimally Important Difference MID

QOL

2. MID Distribution-based methods

Anchor-based methods

3. MID

4. MID COSMIN MID

5. EORTC QLQ-C15-PAL MID

MID

2014/6/21 8 CSP-HOR 37

MID

• MID

• MID

2014/6/21 8 CSP-HOR QOL/PRO 38

2014/6/21 8 CSP-HOR 39

Dr. Wyrwhich