Table 2 Criteria used in overall quality judgement

Origin of items

2 if items were specially collected for use in elderly people with dementia/CI

1 if items were modified for use in elderly people with dementia/CI

0 if items originated from a scale developed for another population

Number of participants*

2 if N => 100 and the number of elderly people with dementia included was considerable relative to the number of items/variables or 50 < N < 100 and corrected for multiple testing

1 if 50 < N < 100 and the number of elderly people with dementia included was considerable relative to the number of items/variables or N < 50 and corrected for multiple testing

0 if N < 50, not corrected for multiple testing and small number of elderly people with dementia included

Content validity

2 if scale seems to cover all important items/dimensions (in the reviewers' opinion): pain items were collected for the specific population and different sources/methods were used to collect items

1 if the scale seems to cover important items/dimensions to a moderate extent (in the reviewers' opinion): items were adapted to the population and different sources/methods were used to collect items

0 if the scale does not seem to cover the important items/dimensions (in the reviewers' opinion)

Criterion validity

2 if correlates acceptable to high (r > .60) according to the 'gold standard' or acceptable according to a 'silver standard' and sensitivity/specificity is determined to be acceptable

1 if correlates moderate-acceptable (.40 < r < .60) according to the 'gold standard' or acceptable according to a 'silver standard'

0 if correlates low (r < .40) or no information is provided

Construct validity in relation to other pain tool

2 if correlates with other pain measures acceptable to high (r > .60)

1 if correlates with other pain measures are moderate (r > .40 < .60)

0 if correlations are low (r <,40) or no information is provided

Construct validity II differentiation

2 if the scale differentiates well (in the reviewers' opinion) between pain and no pain, calm and distressed, pre- and post-medication etc.

1 if the scale differentiates moderately well (in the reviewers' opinion) between pain and no pain, pre- and post-medication

0 if the scale does not differentiate or no information is provided

Homogeneity*

2 if .70 < alpha < .90

1 if alpha > .90 or > .60 alpha <.70

0 if alpha < .60 or no information is provided

Inter-rater reliability # $

2 if reliability coefficient >.80

1 if .60 < reliability coefficient < .80

0 if reliability coefficient < .60 or no information is provided

Intra-rater and/or test-retest reliability $

2 if reliability coefficient >.80

1 if .60 < reliability coefficient < .80

0 if reliability coefficient < .60 or no information is provided

Feasibility

2 if scale is short, manageable with instructions, scoring interpretation

1 if scale is manageable (one format)

0 if scale is more complex