Design
The study is a multi-centre randomized controlled trial with two parallel arms and a 1:1 allocation ratio. One arm receives the intervention plus standard care, and the other arm only receives standard care according to local routines at each ward. In parallel to the intervention, a process evaluation is performed to describe experiences and contextual factors of importance, such as intervention dose and fidelity to the protocol.
Both quantitative data (questionnaires – either online or on paper according to the parent’s preference) and qualitative data (interviews, reflective notes, and audio recordings of the intervention meetings) are currently being collected and analysed. The questionnaires are sent out at baseline (T0), once during the intervention (T1), and at two weeks (T2), two months (T3), six months (T4), and one year (T5) after the intervention. A schedule of the intervention is provided in Table
1.
Table 1
Weekly schedule of the trial
Enrolment | | | | | | | | | | | | | |
Diagnosis | X | | | | | | | | | | | | |
Eligibility screening | | X | | | | | | | | | | | |
Informed consent | | X | | | | | | | | | | | |
Allocation | | | | X | | | | | | | | | |
Assessments | | | | | | | | | | | | | |
Baseline | | | T0 | | | | | | | | | | |
Follow-up measurements | | | | | | | | T1 | | T2 | T3 | T4 | T5 |
Qualitative interviewsa
| | | | | | | Q1 | | | Q2 | | | |
Recording of intervention meetingsa
| | | | | X | X | X | | X | | | | |
Intervention | | | | | | | | | | | | | |
Intervention meetings | | | | | X | X | X | | X | | | | |
Participants
We are currently recruiting participants from two paediatric oncology tertiary care centres in Sweden. These centres cooperate with local hospitals where some of the treatment and supportive care is given; however, diagnoses and major treatments are performed at the centres. Inclusion criteria are being a parent of a child that a) is diagnosed with a first-time occurrence of a malignancy that is curatively treated and b) was diagnosed within the past two months. Furthermore, parents must be able to speak, read, and write Swedish well enough to participate without an interpreter.
All parents, including any stepparents, are being invited to participate, and interventions and measurements are being performed individually. Eligible parents are approached by a recruitment nurse at each site. After receiving information about the study, informed consent is obtained from parents willing to participate and the baseline questionnaire is distributed.
When the baseline questionnaire has been filled in, parents are randomized to one of the two study arms. Parents of the same child are randomized to the same arm to avoid contamination within couples. The allocation sequences were generated using the online service
www.randomization.com. The allocation is performed by an independent person using opaque, pre-numbered envelopes. For each site, two strata are used: i) one participating parent per family and ii) two or more participating parents per family. This strategy will balance the number of parents equally between sites and arms. Both strata are blocked with randomized, varying block sizes (2, 4, or 6 units) that are blinded to the researchers and the person performing the allocation.
Sample size
According to a power calculation, 130 parents in total will be needed if we expect a moderate effect size (Cohen’s d = .5, which corresponds to 13 points of the total score of the primary outcome measure, the Pediatric Inventory for Parents (PIP)), α = .05, and statistical power = .8. We expect that 30 % of parents will not finish the study, meaning that 180 parents in total need to be recruited.
For the qualitative interviews, we plan to interview one third of the parents (n = 20) allocated to the intervention arm.
The intervention
The intervention is based on
the Representational Approach to patient education developed by Donovan and co-workers. This approach combines two theories. The
common sense model of illness representations provides a framework for a thorough assessment of the parents’ knowledge of the topic in question. The mental representation of an illness consists of the following six dimensions: identity, cause, timeline, consequences, cure/control, and emotion. Before giving information, the parents’ representations are assessed by the nurse to find gaps, misunderstandings, or confusion in those representations. This then guides the nurse in tailoring the education session for each individual. To alter a representation, another theory about
conceptual change is used, and this implies that it is crucial to understand the consequences from the knowledge gaps or confusion before providing information to the patient. This method has been used previously in adult services such as oncology and cardiac surgery [
28,
29] and was also shown to be useful in the pilot testing of this intervention [
18].
Two registered children’s nurses at each site have been employed part-time to deliver the intervention. All have specialist training in paediatric nursing and several years of experience in paediatric oncology. Before the intervention started, they took part in a three-day training workshop that included both theoretical sessions and practical training in the intervention method. They were also provided with a booklet outlining the principle content of the method (available from the authors).
Each parent in the intervention arm has four meetings with the intervention nurse, and before each meeting the parent chooses a topic of interest. A list of suggested topics is available for this, based on topics chosen by parents in the pilot study and during a workshop with clinically experienced nurses in paediatric oncology. The parent can also pick a topic on their own or get suggestions for a topic from the intervention nurse. The meeting, which can be face-to-face or over the telephone according to the parents’ preference, is structured around the Representational Approach, that is, it assesses the representations, explores the gaps, errors, and confusions in those representations, introduces new information, sums up the new information, and sets goals. In parallel, parents receive standard care according to local procedures.
The intervention starts two months after the child has an established diagnosis. This point in time was chosen because previous research has shown that both parents and HCPs report that the most intense information provided at diagnosis has usually ceased being given by this time [
4,
6,
30]. The meetings take place at approximately 8, 11, 14, and 22 weeks after the child’s diagnosis.
Outcomes and measurements
The primary outcome for this study is illness-related parental stress as measured by the PIP [
31]. Further, secondary outcomes measured are post-traumatic stress symptoms, anxiety, depression, satisfaction with information, expected and received knowledge, experiences of health care staff, and the number of health care contacts. Background demographic data is also being collected.
The PIP was originally developed in a paediatric oncology setting, but it has been used widely in different paediatric populations. It consists of 42 items that measure both the frequency and difficulty of the following four domains of stress: communication, emotional functioning, medical care, and role functioning. Each item is answered twice on a 5-point Likert-type scale, once fore frequency and once for difficulty. The range for the sum score is 42–210 points, where higher scores indicate higher frequency/difficulty [
31]. The PIP has sound psychometric properties and is available in Swedish.
The
Impact of Event Scale-Revised (IES-R) is used to measure post-traumatic stress symptoms [
32]. The dimensions of intrusion, avoidance, and hyperarousal are measured on a total of 22 items with a 5-point Likert-type scale. The sum score ranges from 0 to 88 points, and higher values indicate more stress symptoms. A Swedish translation is available, and the instrument has been used in paediatric oncology previously.
Anxiety and
depression are measured with two visual-digital scales. VDS-Anxiety has been used previously in this population, and it correlates with other instruments such as Spielberg’s State-trait-anxiety inventory. Parent are asked to assess if they have experienced anxiety during the last week on a 7-point Likert-type scale ranging from never to always. VDS-Depression is the equivalent measure for depression [
33].
Parental satisfaction with information is also measured by visual digital scales. The questions are, for example, “How satisfied are you with the information you have about your child’s illness?” and “How satisfied are you with the information you get from the information sessions?”.
Expected knowledge and received knowledge are measured with two instruments, Knowledge Expectations of Significant Others and Received Knowledge of Significant Others [
34]. Both instruments measure knowledge in 40 areas. Swedish translations are available, and the instruments have been used in several patient education studies.
Experiences with your health care provider are measured on a 15-item instrument developed from Swanson’s theory of caring. This instrument measures aspects of the meeting with the intervention nurse using words such as “comforting”, “informative”, “listening”, and “respectful”. The instrument has previously been translated into Swedish, and psychometric testing of the instrument is in progress.
We also ask the parents for the number of contacts with the health care system and what questions were asked by the parents during these contacts. Summaries of the intervention meetings are written by the intervention nurses after each meeting. These include the topic that was discussed, the duration of the meeting, what components of the representational approach were used, technical problems, and so on.
Parents also report background demographic data such as age, occupation, position in the family, and previous experiences of cancer illness. For the ill child, data about age, sex, and diagnosis are collected.
Data analysis
Quantitative data will be analysed using the SPSS statistical software package. Continuous variables will be reported as the arithmetic mean ± standard deviation, and categorical variables will be reported as the frequency and percentage. Differences between the trial arms will be evaluated using Student’s
t-test for the outcome variables at the available time points. We will also perform regression analyses using data from the process evaluation, including intervention fidelity and other contextual factors. By using multivariate techniques, we will adjust for any differences found in the baseline characteristics of the participants. When appropriate, Bonferroni adjustments will be made on the overall significance levels. For all of these analyses, effect sizes will be calculated. All analyses will employ the intention to treat principle, and two-tailed
p-values ≤ .05 will be considered as statistically significant [
36].
Qualitative data, such as individual and focus groups interviews and recorded intervention meetings, will be analysed using qualitative content analysis [
37]. This method focuses on similarities and differences in data, which is useful for finding possible variances in intervention performance that might explain differences in outcomes.
Ethical considerations
So far, no specific risks with educational interventions such as this have been described. The participants are asked to report any disadvantages to us, and those will be described in the scientific reporting of the study. Participants have the option to withdraw from the study at any time. We will obtain written informed consent by all participating parents. The study was approved by the Ethical Review Board in Umeå (Dnr 2014-167-31 M, valid for recruitment at Umeå and Lund University Hospitals), and is registered at Clinicaltrials.gov (NCT02332226, December 11, 2014).