Article Text
Abstract
Background Nursing home (NH) residents are frequently subject of treatment decisions (TDs) in potentially life-threatening events (PLE). We developed a comprehensive tool to retrospectively evaluate the degree of SDM in PLE, and, in a pilot sample, tested its inter-rater reliability.
Methods Charts of NH residents were screened for pre-defined PLE and related TDs in the past three months. Retrospective evaluation of each of four pre-defined sequential SDM steps was reached by an integrative judgment based on a systematic analysis of four separate sources: (a) fotocopies of relevant passages of resident charts, and audio recordings of short interviews with, as far as available, (b) the resident, (c) the surrogate, and (d) the attending nurse. Altogether 27 such TD sets of fotocopies and audios were presented to five raters, and inter-rater reliability was calculated (SPSS) using Randolph´s κ. Besides SDM, the raters evaluated retrospectively whether care delivered had been consistent with care preferences.
Results Inter-rater reliability of the item care delivered consistent with care preference was good (κ=0.71 [KI:0.55–0.86]); of the items SDM-1 and SDM-2, explaining the situation and discussing options, was acceptable (κ=0.45 [0.28–0.62]; κ=0.43 [0.25–0.62]); and of the items SDM-3 and SDM-4, opportunity to deliberate alternatives and support to reach one’s own decision, was insufficient (κ=0.37 [KI: 0.24–0.50]; κ=0.32 [KI: 0.19–0.45]).
Conclusion Retrospective evaluation of SDM in NH residents is possible using a newly developed tool that bases an integrative judgment on the systematic analysis of four separate sources. This pilot study provides insights regarding possibilities and challenges of the retrospective evaluation of SDM, and regarding possible improvements of the tool.