We asked informal caregivers of patients who visited our outpatient memory clinic between May and August 2019 to participate in a one-time, 3-h focus group. Additionally, we recruited caregivers through our center’s social media accounts. We approached neurologists, geriatricians, nurse specialists and neuropsychologists from various memory clinics in the Netherlands through contacts of the authors and by using a mailing list for members of the Dutch memory clinics network (Nederlands Geheugenpoli Netwerk).
The study was approved by the ethical review board of the VU University Medical Center, and all participants provided written informed consent.
The Amsterdam Instrumental Activities of Daily Living Questionnaire (A-IADL-Q) is an outcome measure that is self-completed by a caregiver and was designed to capture early impairment in daily functioning due to cognitive decline . For the current study, we used the short version of the instrument , which consists of a selection of 30 activities from the original 70-item version. Items were selected based on cross-cultural applicability, frequency of endorsement, and clinical relevance, as judged by clinicians, caregivers and patients . Items are rated on a five-point scale ranging from ‘no difficulty performing the activity’ to ‘unable to perform the activity’. The A-IADL-Q is scored using item response theory (IRT), which accounts for varying ‘difficulty’ of items such that impairment in a more complex activity (e.g., managing the household budget) contributes differently to the total score than impairment in a relatively simple activity (e.g., using the TV remote control). This information is contained in the scoring parameters, as described in detail elsewhere [13, 14]. The total score, or T-score, represents the latent trait of ‘daily functioning’ and is normally-distributed with a mean of 50 and a standard deviation (SD) of 10 in a memory clinic population. Scores thus range from approximately 20–80, with higher scores representing better daily functioning.
We created nine short clinical summaries (‘vignettes’) of fictional patients who had some degree of functional impairment, using combinations of five items of the A-IADL-Q for each vignette. We selected a subset of fifteen items to reduce the number of different activities presented in each vignette and increase comparability between them. The selection was made based on the IRT parameters to have items distributed across the latent trait, so that both more and less impaired ends of the daily functioning spectrum were covered. We then determined what item response category would be most likely to be endorsed given a certain T-score, based on the methods and using an R script adapted from Morgan, Mara . An overview of the most likely item responses of the fifteen items is included in the Additional file 1. The vignettes were created by combining the most likely responses of five items at different T-scores (i.e., different degrees of impairment), and were placed five points (0.5 SD units) apart, ranging from 20 (all most likely item responses were ‘unable to perform’) to 60 (all most likely item responses were ‘no difficulty’). We randomly assigned each vignette a gender, common Dutch surname, random age in the range of 60–70 years, and a stock photo. The vignettes can be found in the Additional file 1.
In the focus groups, we asked each panelist to describe what they considered ‘mild’, ‘moderate’ and ‘severe problems’ in daily functioning, to understand how the panelists defined these categories and create a framework for the subsequent categorization and discussion. Subsequently, panelists individually ordered the vignettes from the one representing the least functional impairment to the one representing the most. Panelists then discussed the order of the vignettes and reached a consensus ordering. Then, panelists individually placed bookmarks between the vignettes to create categories representing no, mild, moderate, and severe problems in daily functioning. This ‘bookmarking’ method was previously developed by Cook and colleagues . Finally, a second group discussion resulted in a consensus categorization. Group discussions were based on the nominal group theory .
Clinicians individually completed an online survey that was modeled after the focus group procedures, and in which they were first asked to describe what they considered ‘mild’, ‘moderate’ and ‘severe problems’. Next, the nine vignettes were presented in order from least to most impaired, and the clinicians were instructed to categorize them into no, mild, moderate and severe problems.
As the clinicians completed the survey independently, consensus between them was determined by taking the mode of the categorization for each vignette (1 = no problems, 2 = mild problems, 3 = moderate problems, 4 = severe problems). The overall consensus categorization was the mode of the three separate consensus categorizations: two from the focus groups with informal caregivers, and the consensus between clinicians. Analyses were performed in R version 4.1.0 .