The investigation of the item distribution of the PCQ-G-S demonstrated a balanced distribution of the six response options. Nine out of 14 items showed an acceptable item difficulty, but five items (item 4, 11, 12, 13, 14) showed a ceiling effect (> 0.8). However, it should not be generally indicated to cancel these items of the PCQ-G-S. Instead, it must be considered that this is an exploratory study and further evaluation of the scale validity are needed including a lager sample. Since the item distribution has not yet been examined in other studies, no comparison of the results is possible. Further research is justified needed here, because the identified ceiling effects affect all items of the subscale climate of community and 36% of all PCQ-G-S items.
The results for structural validity show that the original factor structure of the PCQ-G is robust. Similar to the original Swedish version [13], the 14 items of the PCQ-G-S could be assigned to the three subscales a climate of safety, a climate of everydayness, and a climate of community. The same instrument structure was found for the Swedish (Edvardsson et al., 2009), Norwegian [19], Slovenian [21], and Chinese [20] versions. Thus, Cai et al. (2017) found a stable three-factor solution explaining 73.3% of the total variance for the Chinese version, Bergland et al. (2012) reported the three-factor solution that explained nearly 68% of the variance in the data for the Norwegian version and Vrbnjak et al. (2017) found a three-factor solution that explained 71.22% of the variance in the data of the Slovenian version. The analysis of the original Swedish version resulted in a three-factor solution explaining 60.0% of the total variance [17]. Therefore, scale dimensionality could be seen as confirmed.
Psychometric evaluation of the English PCQ-S resulted in a four-component rotated solution (a climate of safety, a climate of everydayness, a climate of community and a climate of comprehensibility) explaining 71,8% of the total variance [18]. The fourth subscale “a climate of comprehensibility” included four items, relating to the extent staff provided understandable information to patients, patients felt safe, staff were easy to talk to and where patients also had others to talk about their experiences [18]. In the original version, these four items belonged to the subscales a climate of safety and climate of community [17]. Edvardsson et al. (2010) explained the deviation from the original version with three subscales by the fact that the study evaluating the original Swedish version included a sample working on an elective surgery ward with a short length of stay. Because of limited possibility for interactions between staff and patients, the sample in this study may felt prioritising that patients understand implemented medical procedures instead of focusing on proving PCC [18].
Based on a Rasch analysis of the English PCQ-S, residual correlations greater than 0.29 than the mean correlation in the matrix were found. This indicated some evidence of local dependence between two items (item 13 “a place where it is easy for patients to talk to staff”; item 14 “a place where patients have someone to talk”) of subscale three. Since removing or combining item 13 and 14 caused other difficulties, according to Wilberforce et al. (2019) the two items were kept.
The 14-item PCQ-G-S consists of three subscales. It showed strong internal consistency for each of the three subscales a climate of safety (alpha = 0.845), a climate of everydayness (alpha = 0.877), and a climate of community (alpha = 0.867). These results are in line with the results of previous psychometric evaluations. Also, the Swedish [17], English [18], Norwegian [19], Slovenian [21] and Chinese [20] version of the PCQ-S showed internal consistency scores of at least 0.77 for each subscale. Sample sizes in previous studies were comparable to our study. Only in the study of Cai et al. (2017) included more participants (n = 1237).
Strengths And Limitations
A major strength of this study is that a German version of the PCQ is available and additionally, a user manual for the questionnaire (PCQ-G). Thus, an internationally proven questionnaire for the assessment of person-centredness is available for research and practice in the German-speaking countries. Moreover, this is the first study evaluating the psychometric properties of the staff version of the PCQ-G.
This study has some limitations. First, only the staff version of the PCQ-G was evaluated. This means that an evaluation of the patient and family versions is pending and recommended. Second, given the relatively small number of nurses and nursing assistants included in the study, results must be interpreted with caution and have to be proven in a larger study with a confirmatory approach for the PCQ-G-S. Third, the PCQ-G-S was only applied in nursing homes participating in the MoNoPol-Sleep study [22]. Further psychometric validation in different settings is needed to ensure generalisability and to help for further comparisons in different contexts. Moreover, the development of cut-off scores for interpretation purposes is a future goal for the PCQ-G versions.