BACKGROUND AND OBJECTIVES: Given recent advances in psychometric assessment, there is a need for assessment studies using modern test theory in the field of person-centred care, mainly due to the dominant use of analytical strategies based on classical test theory. The main objective of the present study was thus to examine whether selected items from commonly used instruments of person-centred care were able to differentiate between respondents with a reasonably even level of measurement precision across different regions of the construct range using item response theory (IRT).

RESEARCH DESIGN AND METHODS: A Swedish sample of care staff in elderly care (N = 1342) completed a survey including a selection of items from three previously validated measures of person-centred care.

RESULTS: All questionnaire items were submitted to IRT analyses to examine the extent to which the items produced information on the underlying construct. The items exhibited different levels of information. However, in general, for those items exhibiting some information, the pattern of information across the trait range was similar for most of them, that is, the items discriminated better in the lower levels of person-centredness.

DISCUSSION AND IMPLICATIONS: Item response theory analyses are instrumental in creating shorter measurement instruments that may perform nearly as well as the original longer instruments. Given time and other resource constraints in questionnaire administration, there is a gain in only including the most informative items which efficiently and evenly tap the underlying construct along its entire range and in the context of person-centred care assessment this study was an initial step towards this goal. Thus, a set of ten items with satisfactory levels of psychometric quality, that is relatively high information levels across a relatively broad range of the underlying construct, is proposed.