Discussion
Our study is the first to systematically analyze the complex ecosystem in which SDM occurred in a cohort of children with chronic respiratory failure. Prior research on this population has addressed interactions among and opinions of stakeholders in SDM. Hebert et al reviewed 19 transcripts of family-physician conferences regarding tracheostomy placement and concluded that physicians emphasized benefits while minimizing risks14. Nageswaran et al interviewed 56 caregivers of 41 children who received tracheostomies and found that decisional satisfaction was high15. A study that elicited the informational needs of caregivers from 43 families who had faced a decision regarding tracheostomy placement demonstrated that families wanted comprehensive information, including what life would be like at home with a child with a tracheostomy. It also demonstrated that being presented with the option not to perform a tracheostomy was reasonable16. Other studies asked physicians to describe how they go about a shared decision making process17,18.
Two articles allude to how systems in place do not serve the needs of the SDM process18,19. In a Lancet editorial Fine-Gould writes: “For children in ICUs, the health-care professionals involved in making decisions about initiating long-term ventilation are usually not involved in ongoing care once children leave the ICU so a dissociation can appear between the short-term goals of discharge from ICU or hospital and the long-term care goals19.” Large teams with frequently changing members were noted to complicate decision making18,19. Most respondents in a survey of neonatal and pediatric ICU physicians agreed that current ICU systems do not meet the needs of chronically critically ill children18. Fine-Gould recommends reevaluation of the process of initiating long-term ventilation and suggests focusing on the child’s care goals and an ethical framework19.
This current study has several limitations. It is a retrospective chart review with descriptive data. Patients were not screened systematically for the protocol, but rather identified by the members of the working group as they conducted their usual clinical duties. Due to limited available evidence in 2016 when the protocol started, the “12 key questions” were designed by the working group based upon their own clinical experience. Of note, numerous elements of these 12 questions were consistent with recommendations in subsequently published clinical SDM guidelines5,20. Some parts of the protocol may have been accomplished in real time, but not found on chart review and therefore not counted as accomplished.
The patients in this cohort were medically complex, and heterogeneous both medically and psychosocially, which is similar to prior studies of children with CRF. Two prior studies included a greater proportion of patients who received tracheostomy or long-term ventilation15,16. This is attributable to differing methods of patient recruitment. Neither this study nor prior studies prospectively recruited patients from the point where SDM started.
To our knowledge, this is the first study to document the number of services encountered by each child with CRF during a single hospital stay. Our finding that each patient encountered a mean of 17.1 services was striking, but not surprising. Although this number could be overwhelming for any family, counting the number of services grossly underestimates the number of individuals encountered by a patient during a hospital stay. Physician-led services often function in teams comprised of an attending, residents and students, all of whom rotate weekly or monthly. Nurses, respiratory therapists, physicians and other personnel all change shifts at least once per day.
Another striking finding in this study was the challenge of locating documentation of information specific to the needs of SDM. This information was seldom found in one easy to review location, but rather was documented by different providers at different times in different ways. This point is illustrated by the need to eliminate 2 of the 12 key questions from chart review. Overall, the fragmentation and lack of clarity of the information specific to the SDM process found on chart review paralleled the barriers to clear communication of this information in real time among health care professionals.
Taken together, these findings demonstrate a medically complex heterogeneous population cared for by a large number of health care professionals with no clear or standardized means of communicating with each other regarding the information deemed integral to reaching a decision. Such a complex and important decision clearly requires teamwork and “Communication is the essential foundation of an effective team21.” Examination of the PICU through the lens of a microsystem model sheds light on the systems in place which can make it challenging to form an effective SDM team around each child.
A clinical microsystem is defined as “a small group of people (including health professionals and care-receiving patients and families) who work together in a defined setting on a regular basis (or as needed) to create care for discrete subpopulations of patients9.” The anatomy and physiology of a microsystem is often described through the 5 Ps: purpose, patients, professionals, processes, and patterns9. Figure 3 compares the 5Ps of a typical PICU to those observed in 2 of the cases in this cohort. The pediatric intensive care unit has the broad mission of treating acute, life-threatening problems, in a relatively discrete timeframe. By contrast the shared decision making process has the narrower goal of reaching a single decision, typically in a longitudinal manner. Case 1 and case 2 illustrate the heterogeneity of this cohort psychosocially as well as medically. One child initially had no health insurance and hailed from a single parent, non-English speaking family with an unstable living situation. The other child lived in an economically stable home with two English-speaking parents, and consistent health insurance.
The PICU is staffed by a core multidisciplinary team with clearly defined members and roles. Communication patterns are built into daily routines through patient-centered rounds, structured EMR notes, multidisciplinary rounds and sign-outs with shift changes. This pilot program for SDM was undertaken without any change in roles or work-flow. At the end of the first year, the SDM process had only occurred in one case, a problem the investigators attributed to lack of a team leader. Palliative care was then recruited to shepherd the process with a subsequent increase to an average of 9 patients/year for whom the SDM protocol was pursued. Beyond the informal designation of “quarterback” assigned to the palliative care physician, no specific roles were assigned to the health care professionals participating in the SDM process. The limitations of this approach are evident in the finding that the patient decision aide included in the protocol was only given to 7/29 patients.