Study design
A 6 week feasibility study was conducted between October and December 2017 in a nursing home in the North of the Netherlands. The study design was a clustered randomized controlled trial, in which departments of the nursing home were randomly allocated to either the intervention or control group. This study was exempted from METC approval by the METC-WU. Written informed consent was signed by family as legal guardians of participants in this study, since the mental state of the participants (elderly with advanced dementia) made them unable to provide appropriate informed consent.
Study population
From all 24 patients living in the nursing home, 22 patients were included in the study after signing informed consent. These departments were enclosed, i.e. patients who lived in one department were therefore not able to go to the other departments. Exclusion criteria for participation to the study were: having a special diet or receiving enteral tube feeding which does not allow to eat fingerfoods, limited eating options due to chewing and/or swallowing difficulties, and severe illness which makes consumption of fingerfoods impossible for ≥ 2 days.
Intervention
The intervention group consisted of residents of two departments (N = 16). Residents were provided with regular meals, including morning and afternoon snacks as usually provided in the department. In addition, fingerfoods were served daily around 4 pm as snacks, on top of their regular meals and snacks. The control group consisted of residents living in one other department (N = 8). They were provided with regular meals, including morning and afternoon snacks as usually provided in the department, but they did not receive fingerfoods.
Fingerfoods were provided by the general kitchen of the nursing home. The researchers, dietician, and cook of the nursing home developed seven fingerfood recipes: one type of fingerfood for each day of the week. Fruit and vegetable contents of all fingerfoods were around 50% (Table 1). Caregivers provided the fingerfoods daily in two pieces of circa 5x5 cm for each resident, served on small plates. The caregivers encouraged consumption of the fingerfoods, but did not emphasize the health aspects; the fingerfoods were promoted as ‘tasty snack’ instead. If patients needed help with eating, caregivers assisted the patients with eating the fingerfoods.
Outcome variables
Fingerfoods
The caregivers recorded consumption of fingerfoods in the intervention group during the whole intervention period using a checklist. All types of fingerfoods were weighed during both the first and last week of the intervention using a calibrated scale. The average of these measurements was used to indicate portion sizes. Nutritional contents were calculated using Compl-eat (Version 1.0, Wageningen University, Wageningen)(29).
Attitude of caregivers towards fingerfoods
To measure feasibility of the intervention, an evaluation form was provided in week 6 of the intervention to caregivers and other personnel who had been involved in the project. The questionnaire on feasibility consisted of 5 open questions and 3 multiple-choice questions (22).
Fruit and vegetable consumption
Food diaries were completed by the researchers to measure nutritional intake between 2.00 and 8.00 pm before and at the end of the intervention period (t0 and t6), to assess compensation behaviour during diner and to assess total daily fruit and vegetable consumption. The focus of this measurement was on the main meals (served between 5.30 and 6.30 pm), since this was the only meal where possible compensation for the fingerfoods was expected. The residents were randomly divided over three independent weekdays. From the food diaries, daily energy intake (during main meals) and total fruit and vegetable consumption were calculated using Compl-eat (Version 1.0, Wageningen University, Wageningen).
Nutritional status
Nutritional status was measured at t0 and t6 by means of the Short Nutritional Assessment Questionnaire for Residential Care (SNAQrc) (30), body weight, and body mass index (BMI).Body weight in kilograms was measured using a wheelchair scale. Body height was derived from patient information dossiers. To measure change in body weight, body weight measurements were compared with measurements from one month and six months prior to the start of the intervention. BMI was calculated as weight (kg) divided by squared height (m) (31).
Quality of life
To measure quality of life (QoL), the QUALIDEM questionnaire was used (32–34). Caregivers administered the QUALIDEM questionnaire to every resident at t0 and t6. The QUALIDEM questionnaire consists of 37 items which can be divided in 9 subscales on behavioural aspects of every resident during the past week (care relationship, positive affect, negative affect, restless tense behaviour, positive self-mage, social relations, social isolation, feeling at home, and having something to do), rated from never to frequently (score 0 – 3) (33). For each subscale, the sum of scores of all questions was calculated with a higher sum score reflecting a higher QoL. Additionally, the sum of all questions was calculated to give an indication of overall QoL (32,34). The score of all subscales and total sum were linearly adapted to a scale of 0-100 to make the scores more comparable (32).
Other measurements
Baseline information (gender, age, dementia type, use of psychotropic medication, and use of other medication) of all residents was derived from patient information files and recorded anonymously.
Statistical analysis
SPSS (version 22, IBM SPSS Statistics, NY, USA) was used to analyse all data and for all tests, statistical significance was set at P < 0.05. Background characteristics and baseline outcome variables were compared between the control and intervention group. Normally distributed variables were presented as means and standard deviations, non-normally distributed variables were presented as medians and interquartile ranges (IQRs). To test differences of continuous variables between study groups at baseline, independent t-tests or Mann-Whitney-u were used. Categorical variables were presented as numbers and percentages. Chi-square tests were used to test differences of categorical variables between the study groups. If conditions of the Chi-square test were not fulfilled, Fisher exact tests were performed.
The differences in consumption between the types of fingerfoods were assessed with paired t-tests. Energy intake (in kcal) in both study groups during the main meals was compared between t0 and t6 with paired t-tests to check for compensation behaviour after consumption of the fingerfoods.
To measure the influence of fingerfood consumption on daily fruit and vegetable consumption, nutritional status and QoL, analyses of covariance (ANCOVA) were performed. In these analyses, fruit and vegetable consumption (mean daily consumption in grams; t6), nutritional status (BMI; t6), or QoL (overall QUALIDEM score; t6) were added as dependent variable, and condition (intervention or control group) was added as independent variable. Fruit and vegetable consumption (t0), nutritional status (t0) and QoL (t0) were added as covariates.
Paired t-tests were performed to test the effect of the fingerfoods intervention over time in the intervention and control group separately on fruit and vegetable consumption (mean daily consumption in grams; t0 and t6), nutritional status (BMI; t0 and t6) and quality of life (overall QUALIDEM score; t0 and t6). Besides, Chi-square tests or Fisher exact tests were used to measure changes in SNAQrc scores (as indicator for nutritional status).