Next of kin participation in the care of older persons in nursing homes: A pre-post non-randomised educational evaluation, using within-group and individual person-level comparisons

Abstract

Background: Next of kin participation in care is a cornerstone of palliative care and is thus important in nursing homes, and outcomes following interventions need to be evaluated using robust methods.

Objective: To use within-group and within-individual analytical approaches to evaluate the participation of next of kin in care following an intervention and to compare the outcome between the intervention and control groups.

Methods: A pre-post intervention/control group study design was used. The educational intervention, directed towards staff members, focused on palliative care. The Next of Kin Participation in Care scale comprises the Communication and Trust subscale and the Collaboration in Care subscale, with nine items each. In total, 203 persons (intervention group: n = 95; control group: n = 108) were included. Three different analytical approaches were used: 1) traditional within-group comparison of raw ordinal scores and linearly transformed interval scores; 2) modern within-individual (person-level) interval score comparisons; 3) comparisons between the intervention group and control group based on individual person-level outcomes.

Results: Within-group comparisons of change revealed no change in any of the groups, whether based on raw or transformed scores. Despite this, significant improvements at the individual level were found in 32.9% of the intervention group and 11.6% of the control group for the total scale (p = 0.0024), in 25% of the intervention group and 10.5% of the control group for the Communication and Trust subscale (p = 0.0018), and in 31.2% of the intervention group and 10.5% of the control group for the Collaboration in Care subscale (p = 0.0016). However, a significant worsening at the individual level in Collaboration in Care was found in 35.1% of the intervention group but only among 8.4% of the control group (p < 0.0005).

Conclusion: The intervention seems to have a positive impact on next of kin participation in care in nursing homes, especially for communication and trust. However, some next of kin reported decreased participation in care after the intervention. Modern individual person-level approaches for the analysis of intervention outcomes revealed individual significant changes beyond traditional group-level comparisons that would otherwise be hidden. The findings are relevant for future outcome studies and may also necessitate a re-evaluation of previous studies that have not used individual person-level comparisons.

Trial registration: This study is part of the intervention project registered under Clinical Trials Registration NCT02708498.

Conflict of interest statement

The authors have declared that no competing interests exist.

Figures

Fig 1. Individual person-level raw total scores translated to linear logit scale location values (y-axis) for baseline, follow-up, and change in locations (follow-up location − baseline location) (x-axis) together with 95% confidence intervals (CIs).

Panel A illustrates significant improvement (95% CI for change does not overlap zero) in Communication and Trust (CaT) score for Case 147. Panel B illustrates non-significant improvement in Next of Kin Participation in Care (NoK-PiC) score for Case 25. Panel C illustrates no change in CaT score for Case 186. Panel D illustrates non-significant worsening in Collaboration in Care (CiC) score for Case 13. Panel E illustrates significant worsening (95% CI for change does not overlap zero) in CiC score for Case 179.