HSPSC studies

The first European HSPSC studies analysing psychometric properties were conducted in 2005 (Smits et al. 2008 Wenqi 2005). The number of participating hospitals ranges from one to 551 facihties between the studies. Responses in the studies range from 309 to 38,812 participants and from 22 to 92 percent response rates. 

Although we have done our best to establish a sound basis for this overview, the selection presented is neither comprehensive nor exhaustive because it was limited to studies published before October 2013 and responses from national contact points provided to us by December 2013. Thus, we apologise to any researchers or projects that have used the HSPSC outwith those dates in a European country and are not mentioned here. Nevertheless, we hope that this overview will encourage more intense international exchange and collaboration in the future. 

Through the initial web-based literature search on Medline and Web of Science, performed in March 2012, we identified 26 publications on the use of the HSPSC in a European country, which were published between January 2003 (when the HSPSC had originally been developed) and February 2012. We reviewed these publications focusing on instrument development (i.e. the translation and/or adaptation process), data collection process (e.g. in a pilot or field study) and psychometric evaluation (i.e. tests performed and key results of these tests) in different coimtries. 

Overall we identified 44 different European studies that have used the HSPSC questionnaire by Sorra and Nieva (Nieva and Sorra 2003 Sorra and Nieva 2004) as a basis for their survey. These studies were conducted in 20 different Emopean countries. Table 11.1 provides an overview of all reviewed studies using the HSPSC questionnaire in a Enropean country. 

Interestingly, one application of the HSPSC involved an adaptation for the petroleum industry to conduct a comparative study of safety climate differences between healthcare and petroleum industries (Olsen 2010 Olsen and Aase 2010). To oirr knowledge this is the first time that a cross-industry study has used a tool originally developed for the hospital setting. 

Internal consistency was reported in all 14 studies, including results from the original HSPSC version (see Table 11.4). In 13 of the studies, the alphas are based on the original HSPSC structure and one study reported Cronbach s alpha, based on scale refinement in EFA. 

The procedure of psychometric testing varies between the different studies included in this chapter. While many conducted no psychometric tests at all, two studies conducted a confirmatoiy factor analysis (CFA), two performed an exploratory factor analysis (EFA), and 10 studies did both. Despite this difficulty, we found only one study (Piyseley 2008) out of 10 which performed a CFA in order to test the original factor stracture of the HSPSC questionnaire, with several unacceptable thresholds for both absolute and incremental fit indices. With regards to the EFAs, the dimensions Staffing , Communication openness , Organisational learning and Teamwork across hospital units , appeared to be less stable. 

While the focus of almost all studies included in this chapter was the adaptation of a patient safety cUmate measure from the USAto their own national and healthcare contexts, maity of the questions we will have to answer in future might be better addressed at a cross-national level. At the level of the survey instmment this would require a certain amount of consistency regarding the use of terminology and the addition or removal of items - a minimal shared item set. Taking cross-national similarities and differences into account (especially with regard to internal consistencies of the various safety climate dimensions) might help to improve further the overall quality of the HSPSC and to further our understanding of factors at the level of healthcare systems that may have a considerable impact on safety climate. 

Cross-national studies require effective coordination within and between countries. We believe that the first step is a European netwoik of HSPSC users and, to a certain degree, the process of writing this chapter has contributed to such a network by identifying researchers and practitioners with an interest in safety climate assessment. Such a network would also help to connect groups within countries that have developed parallel versions (e.g. Spain and Italy). 

Waterson, P.E., Jackson, J., Stride, C., Hutchinson, A., Hammer, A. and Manser, T. in preparation. The Hospital Survey of Patient Safety Culture (HSPSC) A Cross-national Review of Study Characteristics and Psychometric Properties. Unpublished manuscript. 

---