Skip to main content

Table 1 Missing data by country, and how it was handled in imputation models

From: Multiple imputation of multiple multi-item scales when a full imputation model is infeasible

Variables

% incomplete (for scale variables, % of missing responses to scale items, not scale totals)

Imputed as

Used in imputation models as

Austria

Belgium

England

Germany

Greece

Hungary

Netherlands

Poland

Wales

Total N

323

180

323

274

289

323

237

323

323

  

Gender

0

0

0

0

0

0

0

0

0

1 binary item

Single item

Age

0

0

0

0

0

0

0

0

0

1 continuous item

Single item

Education

3

0

1

0

2

1

0

0

0

1 categorical item

Single item

Marital status

2

0

0

1

1

1

0

1

0

1 binary item

Single item

Employment

2

0

0

1

1

1

0

2

0

1 binary item

Single item

Number of

           

 Medical conditions

1

3

0

1

1

0

0

3

0

1 continuous item

Single item

 Medicines

1

1

0

2

1

1

2

6

1

1 continuous item

Single item

 Tablets

3

3

1

3

1

1

3

12

1

1 continuous item

Single item

 Items prescribed

9

6

3

3

1

8

7

26

6

1 continuous item

Single item

Dosage frequency

1

2

0

0

1

0

1

0

0

1 ordered categorical item

Single item

Morisky adherence

0

0

0

0

0

0

0

0

0

4 binary items

Binary item

Medication adherence rating scale

2

3

1

2

1

2

1

4

0

5 ordinal items

Scale

Prescription payment

1

2

0

1

1

1

*

1

*

1 categorical item

Single item

Affordability problem

1

3

2

1

1

3

1

0

*

1 binary item

Single item

Cost coping strategies

2

7

8

3

2

1

*

8

*

6 ordered categorical items

6 single items

Income

 Source

11

23

22

28

12

8

37

46

15

1 categorical item

Single item

 Perception

8

25

20

28

12

7

37

46

14

2 items: ordered categorical item conditional on binary item

2 items

 Ease of borrowing

9

23

20

28

12

7

38

48

14

2 items: ordered categorical item conditional on binary item

2 items

 Total

9

24

21

28

12

6

38

46

14

2 items: ordered categorical item conditional on binary item

2 items

Health status

1

1

0

0

0

0

1

0

0

1 ordered categorical item

Single item

Practitioner

 Type

7

16

17

23

6

7

29

32

10

1 categorical item

Single item

 Gender

10

17

18

22

14

6

29

38

12

1 binary item

Single item

Satisfaction with

 

 Practitioner

11

18

18

28

2

3

30

35

7

17 ordered categorical items

Scale

 Practice

11

23

18

23

6

3

30

34

11

6 ordered categorical items

Scale

Optimism

6

15

16

20

5

2

26

26

10

6 ordered categorical items

Scale

Illness perception questionnaire: analysed as 8 individual items

9

19

19

25

7

4

31

37

11

8 ordered categorical items

Scale

Necessities

6

15

16

20

5

2

27

27

10

5 ordered categorical items

Scale

Medicine concerns

6

15

16

20

5

2

27

27

10

6 ordered categorical items

Scale

Attitude

9

17

16

24

7

4

30

35

10

7 ordered categorical items

Scale

Barriers (theory of planned behaviour)

10

17

16

23

7

3

30

33

11

1 ordered categorical item

Single item

Facilitators

11

18

16

24

8

5

30

34

11

3 ordered categorical items

Scale

Intention

10

18

17

24

8

4

30

33

10

2 ordered categorical items

Scale

Self efficacy

7

1

16

23

6

3

30

31

10

2 ordered categorical items

Scale

Normative beliefs

10

20

17

24

7

4

30

33

11

3 ordered categorical items

Scale

Barriers

21

22

22

38

18

6

35

35

9

15 ordered categorical items

Scale

Social support

10

19

19

23

20

4

31

43

11

7 ordered categorical items

Scale

Time preference

14

25

23

23

7

21

32

34

19

4 ordered categorical items

4 single items

Discrete choice experiment: not included in analysis

2

5

7

6

2

1

7

12

2

9 binary items

9 single items

  1. Unless stated, variables correspond to a single predictor during analysis. Due to differences in healthcare and prescription systems between countries, not all questions applied to each country. Additionally, in Wales, one question from the barriers scale was not applicable, thus this scale has only 14 items. Whilst illness perception questions were imputed as scale items, they were analysed individually
  2. * Variables not analysed due to differences in prescription policies