Skip to main content

Table 4 Complexity science measures

From: CONNECT for quality: protocol of a cluster randomized controlled trial to improve fall prevention in nursing homes

Concept Measured; Source

Psychometrics; Calculation

Demographics; self-report

Age, sex, job title, years in position, education, and ethnicity (collected at baseline or at enrollment into the study. Categorical measurement)

Communication patterns; all staff

Mean scores on Roberts and O'Reilly openness, accuracy scales [79] and Shortell's timeliness scale [80, 81]. The scales show adequate reliability and validity in various settings [79, 80, 82]. In our preliminary studies scales showed reliability alphas of 0.81, 0.72 and 0.68, respectively; construct validity confirmed by factor analysis and hypothesis testing [17, 30, 83].

NA and LPN participation in decision making about resident care; all staff

Mean score on Anderson et al.'s [83] Participation in Decision-making Instrument (PDMI). The PDMI is established with demonstrated reliability in nursing homes [17, 18, 83–85] and construct validity established through factor analysis [83] and hypothesis testing [17, 84, 86]. Nursing home samples achieved alpha coefficients of > 0.90.

Relational Coordination; all staff

Mean scores on Gittell's [87] five-point scale on which staff will rate interactions between groups (e.g., NA to nurse, NA to dietary; nurse to MD). Three aspects are measured including: frequency of communication; high-quality communication; and supportive relationships [88, 89]. Gittell [88] adapted this scale for nursing homes and achieved a one factor scale and a Chronbach's alpha 0.86. In our preliminary study, we achieved an alpha of 0.95 on a sample of nursing home staff.

Psychological Safety; All staff

Three items from Edmondson's 7-point psychological safety scale that were modified for heath care [90]. The items ask about whether people are comfortable checking with each other or asking questions, whether people value others' unique skills and talents, and whether people are able to bring up problems and tough issues [91]. Studies in healthcare settings reported alphas of 0.74 [91] and 0.73 [92] We slightly revised the scale by changing the word 'unit' to 'nursing home. Because the scale has not been used in a nursing home sample previously we tested the reading level and found that it read at the 6th grade level, which is acceptable for this low literacy sample.

Safety organizing scale; all staff participants

Mean score on Vogus and Sutcliff's scale designed to measure five 'interrelated behavioral processes: preoccupation with failure, reluctance to simplify interpretations, sensitivity to operations, commitment to resilience, and deference to expertise' [93, p. 47]. In a large sample of hospital RNs, the 9-item, 7-point scale showed reliability (alpha = 0.88), convergent and discriminant validity, and criterion validity, and was reliability aggregated to reflect a unit-level construct [93]. We revised the wording for nursing homes. Alphas were > 0.90 in both the baseline and follow up survey in our preliminary studies.