Consensus on interprofessional collaboration in hospitals: statistical agreement of ratings from ethnographic...

Few methods are available for analysing psychometric properties of combined qualitative and quantitative data. While conventional reliability of measures - meaning reproducibility or consistency - may not be meaningful in small-N research, in some health services studies agreement on perceptions arising from data generated by fieldwork and quantitative measures can be examined to good effect.

We studied interprofessional collaboration (IPC) in seven hospitals. An ethnographer shadowed and conducted interviews with regulated health professionals in medicine wards. Concurrently, nurses completed the nurse-doctor relations subscale of the Nursing Work Index (NWI-NDRS) and a new measurement scale for IPC with doctors in the domains of communication, accommodation, and isolation. After fieldwork, the ethnographer rank-ordered hospital sites on IPC from 1 to 7 based on interpretation of the qualitative data. Mean-scale scores were calculated for hospital sites and converted to ranks similarly. The Tinsley-Weiss T-index (Tinsley & Weiss, 1975) for agreement among rank orderings was calculated for dyadic combinations of fieldwork and measurement ranks.

Perfect agreement was obtained for the most liberal agreement definitions considered - differences of two rank positions - involving qualitative data agreement with IPC subscales for accommodation and isolation. Defining agreement as a difference of 1 rank at most, the T-index was 0.77 for agreement between fieldworker and IPC accommodation and the same for NWI-NDRS and IPC isolation.

Qualitative data from fieldwork rankings were substantially in accord with the contemporary IPC scales, less so with the NWI-NDRS. Qualitative data appear to be useful as an additional approach to confirming the validity of quantitative scale data in measuring a complex interpersonal relational construct.