Adaptation, data quality and confirmatory factor analysis of the Danish version of the PACIC questionnaire

Helle Terkildsen Maindal, Ineta Sokolowski, Peter Vedsted

34 Citationer (Scopus)

Abstract

BACKGROUND: The Patient Assessment of Chronic Illness Care (PACIC) 20-item questionnaire measures how chronic care patients perceive their involvement in care. We aimed to adapt the measure into Danish and to assess data quality, internal consistency and the proposed factorial structure.

METHODS: The PACIC was translated by a standardised forward-backward procedure, and filled in by 560 patients receiving type 2 diabetes care. Data quality was assessed by mean, median, item response, missing values, floor and ceiling effects, internal consistency (Cronbach's α and average inter-item correlation), item-rest correlations and factorial structure was assessed by confirmatory factor analysis (CFA).

RESULTS: The item response was high (missing answers: 0.5-2.9%). Floor effect was 2.7-69.2%, above 15% for 17 items. Ceiling effect was 4.0-40.4%, above 15% for 12 items. The subscales had average inter-item correlations over 0.30 and CFA showed high factor loadings (range 0.67-0.77). All had α over 0.7 and included items with both high and low loadings. The CFA model fit was good for two indices out of six (TLI and SRMR).

CONCLUSIONS: Danish PACIC is now available and validated in primary care in a type 2 diabetes population. The psychometric properties were satisfactory apart from ceiling and floor effects. We endorse the proposed five scale structure. All the subscales showed good model fit, and may be used for separate sum scores.

OriginalsprogEngelsk
TidsskriftEuropean Journal of Public Health
Vol/bind22
Udgave nummer1
Sider (fra-til)31-6
Antal sider6
ISSN1101-1262
DOI
StatusUdgivet - feb. 2012
Udgivet eksterntJa

Fingeraftryk

Dyk ned i forskningsemnerne om 'Adaptation, data quality and confirmatory factor analysis of the Danish version of the PACIC questionnaire'. Sammen danner de et unikt fingeraftryk.

Citationsformater