Advance Care Planning in Nursing Homes in Flanders

Study hypothesis The primary study hypothesis is that the introduction of the ACP+ programme in nursing homes will improve the knowledge and self-efficacy of nursing home care staff regarding advance care planning (ACP).

Ethics approval The study was approved by the Ethics committee of University Hospital Brussels (Vrije University Brussels, 22/02/2018, ref: 18-003 - B.U.N. 143201834759)

Study design A cluster randomized controlled trial using baseline and follow-up measurement of relevant outcome variables and process evaluation. The trial will be implemented in a sample of 14 nursing homes (see eligibility criteria below) randomised to either intervention or control group.

Trial duration The duration of the entire trial, from the moment the first nursing home is recruited until the end of follow up will be 10 months.

At month 0, immediately after randomization, the baseline assessment is completed in both intervention and control homes. Staff did not know the outcome of the randomisation when filling in the baseline measures (with the exception of the nursing home director).

At month 1, the ACP+ programme will start in nursing homes in the intervention group. This programme consists of two parts, each lasting 4-months (see intervention below for more details), taking up a total of 8 months.

At the end of month 8, follow-up measures will be collected in both groups. The process evaluation will run through the intervention period, starting at month 1 and will be completed at month 9.

Randomisation Paired randomization of nursing homes to the control and intervention groups is performed by an independent and blinded statistician. Criterion is facility status (public vs. private without profit objective vs. private with profit objective).

Sample size calculation Given an expected effect size of d=0.50, the analysis will have power of 80.27% to detect a difference between intervention and control group on the primary outcome (knowledge and self-efficacy of care staff regarding ACP) at follow-up compared to baseline, with a two-sided significance level α=0.025.

Assuming an intra cluster correlation coefficient (ICC) of 0.036, we need to involve at least 161 staff members per study arm. To allow for an initial response rate of 70% and a staff turn-over of 10%, this number is increased to 242 staff members per study arm, 484 members in total. As the average staffing level is 5 FTE for nurses (including 1 FTE head nurse) and 5 FTE for care assistants per 30 beds and the included nursing homes are all required to have over 100 beds, we expect at least 35 care staff members per cluster (nursing home). We will, therefore, need to include 14 nursing homes to achieve the necessary sample of care staff.

Intervention The ACP+ programme is offered in addition to any standard education or continuous training that staff receive in each of the participating nursing homes randomised to the intervention group. Throughout the ACP+ programme, a professional trainer is made available by the research team to train the nursing home staff and to guide implementation of the programme. The ACP+ programme aims to improve or establish advance care planning (ACP) in the day-to-day routine of staff working in nursing homes. At the core of the intervention is the nomination of several representatives for ACP (named ACP reference persons) in each facility who are trained by a specialised ACP trainer made available through the study. During the study period, the specialised trainer will provide training and support to all staff in the nursing homes and specifically support the ACP reference persons in developing their knowledge and skills regarding ACP, implementing the ACP+ programme and training other nursing home staff.

The intervention consists of two parts. The first part focuses on buy-in from management and development of facility-level ACP policy and training staff. A central element is the training of 'ACP reference persons' (who are nurses of the facilities, 3-4 per facility), who will themselves train the other staff in ACP, with the support of the external trainer (train-the-trainer approach). The second part focuses on nursing home staff conducting ACP conversations with residents and family. In addition, the trained staff will be offered more in-depth training sessions (so called 'comeback-seminar' to evaluate the progress they have made so far, and 'specialisation sessions' on handling ACP with residents living with dementia and professional communication) to extend their knowledge and skills, as well as one-on-one coaching.

The intervention implementation period has a total duration of 8 months and is divided into:

• a four-month preparation and training phase. During this phase the ACP reference persons will attend a "two-day training" given by the ACP trainers. Other staff will receive training by the reference persons on conducting ACP conversation or recognizing triggers for an ACP conversation in nursing home residents.
• A four-month follow-up phase in which ACP conversations are held with residents. Additional training sessions will be organized to give more in-depth knowledge to the ACP reference persons.

The staff of nursing homes in the control group will receive no additional training next to any standard education or continuous training. After the intervention and follow-up measures are finished, all nursing homes in the control group will be offered a shortened version of the ACP+ training programme as well as all ACP+ training materials.

Statistical analysis The researchers will conduct regression analyses to test study hypotheses using intention-to-treat (ITT) and per protocol approaches. Multi-level mixed model regression analyses will be used to analyze outcomes, controlling for baseline values and clustering of data within the nursing homes. Within the multi-level mixed model analyses, outcomes will be analyzed with nursing homes as random factor, and group, time point, and their interaction as fixed factors. Differences in mean change between the intervention group and the control group (group by time interaction) will be calculated. The effect sizes (Cohen's d) using the baseline-adjusted mean differences and the variance between nursing home staff and between nursing homes will be estimated to assess effect sizes for the different outcome measures.