Methods
The purposive sample of participants was recruited through organizations that care for families with children receiving PPC. The PPC team at Starship Children's Hospital was the first organization contacted by the researcher. Team members oversee the provision of care to PPC patients and their families in Auckland and provide advice to the rest of New Zealand. The PPC team provided the researcher with a letter of support and agreed to serve as the link between the researcher and potential participants. Three other organizations were identified by the PPC team to be consulted: the Child Cancer Foundation and Heart Children Foundation in Auckland and "True Colours" in Hamilton, New Zealand. Family support or other professionals (approximately 12 individuals) from the 4 organizations agreed to assist with the recruitment of participants between February 2010 and February 2011. Introduction to the research by these trusted individuals and organizations undoubtedly helped to circumvent perceived barriers associated with potential "gatekeepers" (parents, clinicians and others who cared for the children).
The interview (Table 1) was designed with the help of the Starship PPC team and Maori and Asian advisors. It aimed to lead parents gently into the topic of discussing/concealing the impending death of the child with the PPC patient and his/her siblings. Insights into the perceptions of death that caregivers believed their children held and how they communicated were sought as well as advice on approaching families for future research.
The aims of the research were given in the information sheet to ensure cultural sensitivity and to respect ethical guidelines. They were notified of the study as their child was receiving palliative care or approximately 6 months or more after the death of their child, as suggested by a study of bereaved parents following the death of their child. Ethical approval was granted by the Northern X Regional Ethics Committee, reference no. 4529. No competing financial interests exist.
Adults who considered themselves "primary caregivers" of verbal children receiving PPC or who had received PPC prior to their death were invited to participate. The majority of the families were approached first by the PPC team, who identified them as English-speaking and within an appropriate age range and cognitive ability. Over the course of the year, the principal investigator (PI) received consent from the organizations to approach 15 families and attempted initial contact via phone calls. Two of the families were unresponsive to the initial calls. The PI met with the remaining 13 families (21 caregivers) who consented to having the study explained to them. Nineteen of the 21 caregivers approached by the PI agreed to take part in the study. Data collection ceased once the data were perceived as having reached a point of saturation.
The recruiting team counselors (mostly from the PPC team) spoke directly about the research with the families and delivered brochures outlining the intent and aims of the study to them. The researcher met with the primary caregivers at their homes to explain the rationale and the procedural elements of the study and gain consent.
Consenting participants were interviewed in their homes or workplaces according to convenience. Interviews, which lasted 25 to 93 minutes, were audio recorded. Once collected, the interviews were transcribed and analyzed (Table 2) by the PI, discounting differences between the bereaved and nonbereaved as well as male and female data sets and applying Braun and Clarke's method of inductive thematic analysis. The transcripts were stored in NVivo (QSR International, Melbourne, Australia), a computer program for qualitative analysis.
The transcribed interviews were read multiple times, and initial ideas for topics commonly repeated within and between them were recorded. Quotations that related to the research topic (communicating about the patients' deaths or avoiding communication) were coded on a semantic level without interpreting the caregivers' statements. A few of the same quotations were allocated different codes (eg, "talking about death" or "avoiding talking about death"). The coding procedure is outlined in Table 2. The researcher repeatedly asked, "Why did [caregiver] speak/avoid speaking with [child] about their impending death?" Once analyzed without discrimination, the sex and bereavement status were noted in this article to clarify contextual factors and ensure the transparency and credibility of the data.