Aims
The effect evaluation provides insight into the effectiveness of the intervention on beliefs regarding organ donation and the intention to register (as donor or non-donor).
Design
This article adhered to Consolidated Standards of Reporting Trials (CONSORT) guidelines (see Additional file1 and Figure 1).. The effectiveness of the intervention was investigated in a post-test only cluster randomized controlled trial, in which the intervention was offered to the experimental group and after measurement also to the control group. This design was most suitable in a school setting to ensure that all students would receive the intervention. Participants in the control group filled in the evaluation questionnaire just before receiving the intervention, while participants in the experimental groups did this at the end of the second lesson. Randomization to the experimental and control groups took place at the class level. More details on the content of the intervention, study design and methodology can be found in the protocol article[12].
Recruitment & procedure
Recruitment started in September 2017 and stopped in June 2018, as the summer vacation started in July. Teachers of Dutch schools for IVE who taught a course on Citizenship were invited by e-mail to use the intervention and participate in the study. Region and study discipline (Health and Well-Being, Engineering, Economic, Social Services, Green and Security) were taken into account to ensure a good representation of the target group. Students from level 2 to 4 were included. No additional inclusion or exclusion criteria were set, as students were recruited via their teachers. Students were typically in the age of 15 till early 20s, but older students were no exception nor excluded.
Teachers who accepted the invitation were asked to draw up an inventory of the number of classes for participation, either being taught by themselves or colleagues. This led to the inclusion of 132 classes in 12 schools. Classes were matched in terms of school, study discipline and educational level and then randomly assigned to either the control or experimental condition by computer software (N = 66 classes in the experimental group, N = 66 classes in the control group) (Figure 1)..
Teachers received a personalized link via e-mail to reach the website with the educational program. This link was used by both the teacher and the students of that class. Further, they received the questionnaires, instructions and a manual on paper. Teachers were instructed to either distribute the questionnaires prior to delivering the intervention (control groups) or the other way around (experimental groups). The control classes completed the questionnaires before the first lesson. All participants then received the intervention, taught by their own teachers. The intervention consisted of two 50-minute lessons addressing three elements. The first lesson focused on increasing involvement, encouraging positive beliefs and counterbalancing negative beliefs. This was done by watching video fragments followed by plenary discussions. In the second lesson students first received tailored feedback on misconceptions they might have had by individually working with two quizzes on the website. Moreover, they did an exercise on filling in a simulated organ donation registration form to enhance their self-efficacy. Immediately after the intervention was delivered, students in the experimental classes filled out the questionnaires.
Teachers were asked to return the forms afterwards. Not all forms were returned, leading to a drop-out of N = 44 classes, of which one entire school (with N = 22 classes) (Figure 1).. Drop-out was due to problems with planning the lessons or staff turnover, leaving N = 88 classes (N = 45 in the experimental group, 617 students, N = 43 in the control group, 553 students).
Six months to one year after implementation, teachers were contacted for a follow-up measurement. This entailed an online questionnaire for students who received the intervention. We decided not to recruit new control groups (while this was indicated in the protocol article[12]), as this appeared unfeasible. Instead, we compared registration rates of the follow-up measurement with the original control group. The follow-up measurement led to a response of N = 189 students. High drop-out rates were found because of multiple reasons such as students graduating the year before, being spread over new classes, and conducting an internship (and therefore not being present at school).
Ethical approval
The study was approved by the Ethics Committee of the Faculty of Health, Medicine and Life Sciences, Maastricht University, on 23 October 2017 (reference number: Steenaart/231017) and registered at the Dutch Trial Register (NTR6771; https://www.trialregister.nl/trial/6557). Students provided informed consent after explanations about the study, the opportunity to withdraw their consent at any time, and the anonymous way in which their data would be processed. They were provided with contact information in case they had any further questions.
Measurements
The questionnaire in this study was largely based on the questionnaire used in a preparatory study[7]. The questionnaire assessed demographics, registration behavior, registration intention and beliefs regarding organ donation. The full questionnaire is available at https://osf.io/2jf7v/?view_only = 4ed46d09a1874184a17636ebbd019415.
Demographics. Demographic variables included sex, age, study discipline, study level, country of origin and religion. Country of origin of the student, mother and father were assessed and combined into one item; having a western or non-western background. Students for whom one or more parents were from a non-western country or who were born in a non-western country themselves, were labeled as non-western, while students from whom both parents were from a western country and who were born in a western country themselves were labeled as western[13]. Religion was also dichotomized into being religious or not.
Registration behavior and intention. Participants were asked whether they already registered a decision regarding organ donation (yes/no/don’t know) and, if applicable, what choice they registered (posthumous organ and tissue donor/posthumous donor for specific organs and tissues/non-donor/ leave the decision to the next of kin/leave the decision to a specific person). Unregistered participants were then asked whether they intended to register a decision (yes/no/don’t know) and, if applicable, the choice they would like to register (see above mentioned options). The intention to register was the primary outcome measurement. All registration related questions were dichotomized. Registration status was dichotomized into registered (yes) and not registered (no/don’t know), while registration choice was dichotomized into being a donor (posthumous organ and tissue donor/posthumous donor for specific organs and tissues) and not being a donor (non-donor/ leave the decision to the next of kin or leave the decision to a specific person/don’t know). The intention questions were dichotomized in the same manner.
Beliefs regarding organ donation. The beliefs regarding organ donation included items related to attitude, self-efficacy, knowledge and social outcomes. The questionnaire contained 25 questions on beliefs, all answered on a 7-point Likert scale (totally disagree to totally agree). These questions were based on constructs from over 30 (mainly qualitative) existing studies and have been used before[7].
The one-year follow-up measurement consisted of a short online questionnaire, primarily assessing registration status (yes/no/I don’t know) and if applicable, the choice they registered (see earlier mentioned options). However, as the response rate at follow-up was very low (N = 189) and over 50% of these participants did not remember receiving the program, the results of this follow-up measurement will not be reported, as no valid conclusions can be drawn from this.
Statistical analyses
All data were analysed using SPSS 24 and.05 was used as significance level for all analyses. The first aim of this study was to investigate the effectiveness of the intervention on the intention to register a decision regarding organ donation (i.e. the primary outcome in the study at hand). These analyses were only done among people who were not registered yet or did not remember (N = 916) (Figure 2).. Secondary outcomes were the choice students intend to make and beliefs regarding organ donation. Since students were nested within classes within schools, multilevel analyses were used to examine the relationship between the intervention and (determinants of) organ donation registration intentions.
The effects of the intervention on the primary outcome were examined using multilevel nominal regression analyses, in which intervention group (control/experimental) was the main independent variable. Two different analyses were run; one for the dichotomized outcome (yes/no) and one for the original outcome (yes/no/don’t know). Both analyses were adjusted for several demographic characteristics (i.e., sex, age, educational level, religion, migration background and having had organ donation education before) as they were expected to be associated with the outcomes. To identify possible interaction effects, interaction terms of age, sex and educational level with the intervention variable were separately added to the analysis model in case of the dichotomized outcome (as this was our primary focus).
The effects of the intervention on the choice students intended to make was done in a similar analysis. However, only the dichotomized outcome (donor/non-donor) was used, as some categories of the original outcome were chosen by too few students (e.g. I want to leave the decision to a specific person was chosen by only 3 students) to allow for reliable analyses. The model was adjusted for the same demographic characteristics as were used for the primary outcome.
Differences in beliefs regarding organ donation were examined with multilevel linear regression analyses, in which the intervention group (control/experimental) was again the main independent variable. A total of 25 analyses were run, one for each belief separately. These models were again adjusted for the same set of demographic characteristics. Seven beliefs were strongly skewed to the right. For these beliefs, a gamma regression was done. To control for multiple testing, the Benjamini-Hochberg procedure was used with a false discovery rate of 5%[14]. An excel template was used to calculate the adjusted significance levels.
The data file, syntax and excel templates used are available at https://osf.io/2jf7v/?view_only = 4ed46d09a1874184a17636ebbd019415. These efforts are taken to maximize scrutiny, foster accurate replication, and facilitate future data syntheses (e.g., meta-analyses)[15].