The preliminary measurement properties of the person-centred community care inventory (PERCCI)

Mark Wilberforce, David Challis, Linda Davies, Michael P. Kelly, Chris Roberts

Researchers investigating person-centredness in older people's long-term community care are hindered by the lack of appropriate measures. Studies have tended to rely on proxy indicators or generic instruments, risking invalid results. This new research aimed to develop and psychometrically test a person-centredness scale for use in older people's community services. Questionnaire items were sourced from groups of older people and mapped to a conceptual framework of person-centredness. A postal questionnaire in 2015-2016 tested these items with older people supported by mental health and social care services in five areas of England. Dimensionality was assessed through exploratory factor analysis and a confirmatory bifactor model, with classical item analysis removing weak items. Test-retest analysis was undertaken through a repeated postal questionnaire 3 weeks after the first. Three factors were identified, representing (i) interpersonal and (ii) organisational aspects of person-centred care; and (iii) negatively phrased items. Removing weaker items resulted in an 18-item scale. The bifactor analysis concluded the summary scale was 'essentially unidimensional'. The Person-centred community care inventory (PERCCI) had excellent reliability, with Intra-Class Correlation Coefficient of 0.886 [95% CI 0.818-0.929]. A priori hypotheses about associations with satisfaction metrics and support variables were broadly confirmed. The PERCCI has promising measurement properties and can be recommended for use in research with older adults using community mental health and social care services. Future developments must identify how sensitive the instrument is in detecting changing service quality.