Modeling and Evaluating Evidence Based Continuing Education in Nursing Home Dementia Care

Purpose:

Dementia is a major challenge for our society. In Norway, around 40 % of persons suffering from dementia live in nursing homes (NH), and the majority experiences severe functional impairments and complex needs including associated behavioral changes. International studies indicate that NH residents with cognitive impairment and high dependency are most likely to be restrained physically. There is evidence of various adverse effects such as injuries, reduced psychological well-being or decreased mobility related to the use of physical restraint. Moreover, in addition to ethical and professional arguments, there is evidence supporting that user-involvement and shared decision-making have positive effects for both residents and staff. The aim of the present study (MEDCED) is to carry out an educational intervention in care staff to test its potential effect on use of restraint in NH residents with dementia and increase job-satisfaction in care staff.

Description:

This is a single-blinded controlled cluster-randomized intervention trial recruiting staff from 24 NHs in three counties in the region of Western Norway, randomly selected from the total NH population in the region. A cluster is defined as a working ward of a larger NH, without shared facilities or staff. The 24 NHs are randomized into two groups, for intervention (12) and control (12). One department in each is selected for intervention and data collection will take place also in one department in each of the NHs in the control group. Hence the analyzed units (clusters) are NH departments. The educational intervention (EI), Relation Related Care (RRC), consists of a two-day seminar and one hour monthly guidance for six months. A seven-month educational intervention will be performed, and due to ethical issues the control group will receive the same intervention after the second follow-up. Effects will be measured on resident and care staff level at baseline before randomization, after the intervention (7 months) and at follow-up (7 months later) of the intervention group.

An analysis of how many units needs to be included are based on 12+12 clusters. Two calculations have been performed. The first calculation is based on how big is the reduction in use of restraints. We assume in our model that there will be a 60 % reduction in use of restraints in the intervention group and 33 % reduction in the control group (Hawthorne effect). For this reason we need to include averagely 13 patients in each department/unit in order to achieve a statistical power of 0,8.