Participants and procedures
Over a 4-year period starting July 2016, a SDM process (Figure 1) was implemented in routine care of patients with chronic respiratory failure (CRF) who were < 21 years old, admitted to an inpatient pediatric service, and for whom tracheostomy was a consideration. Eligible patients were identified by the investigators as part of usual care, without implementation of a specific screening procedure. Initially two core elements comprised the protocol, the first of which included sharing educational material with the family. This included a patient decision aide called The Child Tracheostomy Decision Guide10 created by Winnipeg Regional Health Authority, which meets most of the standards set by the International Patient Decision Aid Standards Collaboration11. The second core element was comprised of 12 key questions designed to elucidate medical and psychosocial information which the working group deemed integral to making an informed decision. The SDM protocol was only implemented in one patient during the first year, after which time a third core element was added to the protocol- consultation of the pediatric palliative care team.
Each of the 3 investigators independently performed retrospective chart reviews on 1/3 of the 29 patients deemed evaluable. Data collection focused on the three core elements of a medical microsystem: the patient, the provider and information9. After determining in which hospital admission the decision occurred, the investigators extracted data using detailed questions in a REDCap database. Finding much of the data, such as medical diagnoses and information on the SDM process, required the medical expertise and judgement of the investigators, all of whom were fellowship trained pediatric physicians. To address concerns regarding inter-reviewer reliability, a second chart review was performed on each patient by a different investigator, without knowledge of the previous data collection. Investigators then compared the two analyses for congruence and reconciled any discrepancies by consensus re-review of the medical record. During the re-review process, there were two inter-reviewer differences regarding which admission to review, both of which were resolved by consensus opinion.