Measuring the level of implementation of advance care planning - a fidelity-based cross-sectional study.

Introduction: Advance care planning (ACP) is supported by evidence, legal and ethical principles and ACP guidelines. However, this complex intervention is often poorly implemented. This article presents a novel fidelity scale to measure the implementation of ACP and reports the preliminary testing of the scale at baseline in a larger trial aiming at implementing ACP in hospitals in Norway. Method: The research design conducted was a cross-sectional measurement of fidelity to guidelines, conducted in 12 geriatric hospital units in Norway recruited using purposive sampling. The sample size for the larger trial was determined, based on prior research, to be at least four clusters in each arm. Due to the possibility of unit drop-out during the project period and to secure sufficient power, we included six units in the intervention arm and six in the control arm. For the baseline measurement we developed an ACP fidelity scale including three subscales: implementation, quality and penetration rate, each with 4-11 items. We ensured face and content validity through input from relevant theories and research, experts and users. Items were scored from 1 to 5, with 1 indicating no implementation and 5 indicating full implementation. Data was analyzed using descriptive statistics, Cronbach's alpha and calculation of interrater reliability for the scales. Further psychometric testing could not be conducted at this stage due to the lack of variance in the items. Results: The mean score for the implementation subscale was 1.213, with a median of 1, a standard deviation (SD) of 0.08, a standard error (SE) of 0.01, a variance of 0.01, and a range of 0.28 (minimum 1.14 and maximum 1.42). The scores in the subscale showed that none of the units had recommended implementation measures. Only a few professionals reported they had heard of ACP, but not the whole staff. Cronbach's alpha could not be estimated due to the lack of variation in the scores for this subscale. On the quality subscale, which assesses whether ACP is practiced in accordance with practice guidelines, the mean score was 1.11, the median was 1, the SD was 0.48, the SE was 0.06, the variance was 0.13, and the range was 1.27 (minimum 1 and maximum 2.27). The scores in this subscale showed that ACP was practiced sporadically by the palliative care team in only one unit, while the other staff did not engage in this practice at all. Cronbach's alpha for the subscale on quality was 0.887 (11 items) showing an acceptable internal consistency. For the penetration rate subscale, which measures how widespread the practice is, the mean score was 1.08, the median was 1, the SD was 0.28, the SE was 0.05, the variance was 0.08, and the range was from a minimum of 1 to a maximum of 2. Among the total number of admitted geriatric patients, only 10% had received ACP in only one of the 12 units. Also, for this subscale, the model for Cronbach's alpha could not be applied. There was little variation in the low measurements, thus the interrater reliability was high, reflected in the Intraclass Correlation Coefficient (ICC). The ICC was 0,916 [-0,721,0.976] for the implementation subscale, 1.00 for the quality subscale, and 1.00 for the penetration rate subscale. Conclusion: Our findings indicate very low implementation of ACP in acute geriatric hospital units in Norway. The newly developed ACP fidelity scale has the potential to serve as an important tool for improving the quality of healthcare services for older patients. However, more data is needed to validate the psychometric properties of the scale. Our study should be considered as a preliminary study, and the scale should be used with caution as long as its properties are not well validated. Clinical Trial Registration: ClinicalTrials.gov, NCT05681585.