A Rasch Analysis of the Person-Centred Community Care Inventory (PERCCI): Older People’s Experiences of Long-Term Community Support in England.

Abstract Background Person-centredness has become the byword for a modern, high quality care service. Yet measurement of person-centred quality is fraught with difficulty, and existing instruments tend to rely on classical psychometric methods with many untestable assumptions. An important new instrument, the Person-Centred Community Care Inventory, has yet to be subject to modern psychometric techniques.Methods The same 22 initial items used to form the Person-Centred Community Care Inventory were re-examined under a Rasch framework. A re-analysis of nearly 600 questionnaire returns from older people in England with community care needs was undertaken, with this large sample partitioned into a ‘evaluation’ and ‘validation’ sample. In addition, a test-retest inspection of 77 repeat administrations was undertaken.Results Rasch analysis confirmed a more parsimonious form of the scale, comprising two subscales which broadly met key diagnostic tests. Six items evaluated person-centredness in care worker interactions, and a second six items evaluated person-centredness in care plans. When combined into a single 12 item scale using ‘testlets’, the resulting PERCCI-12 suggested satisfactory measurement properties overall, albeit with reduced internal reliability over two separate subscales. Test-retest reliability of the PERCCI-12 was estimated at 0.871 and was robust to sensitivity tests.Conclusion The Rasch evaluation of the PERCCI supports a shorter scale formed of 12 items. This can be analysed as two separate subscales, or as a single summary scale with weaker internal reliability. Future research is needed to reinforce construct validity, to improve interpretability, and to establish minimally important difference.