Caregiver Decisional Conflict Before and After Consultation About Gastrostomy Tube Placement

We conducted the project at a freestanding tertiary care children’s hospital in Canada that manages ∼16 000 admissions per year. Approximately 200 enterostomy tubes are inserted annually at our institution. Most tubes are inserted by interventional radiologists using image-guided percutaneous techniques; surgically placed tubes are reserved for specific indications, such as failure of percutaneous techniques or for children with anatomic abnormalities such as tracheoesophageal fistula.

The patient population included all children referred for image-guided g-tubes who received ambulatory or inpatient g-tube consults with their primary caregiver to discuss initial g-tube placement. Children were excluded if (1) their primary caregiver was non-English speaking and reading, (2) they were being considered for gastrojejunostomy tube placement because different risks and benefits may influence decisional conflict, (3) they had a projected life expectancy of <2 years, (4) the primary decision-maker was not the primary caregiver (eg, children in custody of child protection services), or (5) study processes conflicted with patient care (eg, families who were late for clinic or children who were determined not to be g-tube candidates). The project ran from April 4, 2017, to September 7, 2018.

Since 1999, the hospital has held a weekly half-day clinic to assess children referred for g-tube internally by specialists and externally by community-based providers. A dedicated pediatric nurse practitioner, a dietician (if a dietician is not already involved), and a pediatric hospitalist have an appointment of 1.5 to 2 hours with the family. The appointment discussion addresses the following questions. (1) Is the g-tube indicated? (2) Is it technically feasible? (3) Are there safety issues for g-tube placement? (4) Is g-tube placement aligned with the family’s values? At the end of the clinic visit, families of children who are determined to be good candidates can choose whether to schedule tube placement, decline tube placement, or delay deciding. Caregivers who choose to proceed are scheduled for a pre–tube placement hands-on g-tube class run by the nurse practitioner and a g-tube resource nurse. Undecided families have the option to attend the class or have additional conversations with the hospitalist and nurse practitioner. On the day of the procedure, the interventional radiologist meets with the family to explain the procedure again and to obtain consent before proceeding with g-tube placement. Approximately 120 children are evaluated per year in g-tube outpatient clinic, and ∼60% opt to proceed with tube placement.

Children who are referred for inpatient g-tube placement by interventional radiology receive similar evaluation, counseling, and decisional support during an in-hospital consultation by the same team. The process is the same for all referring teams, including the NICU. Approximately 150 inpatients are evaluated per year, and ∼85% opt to proceed with tube placement.

The Building Excellence in Enteral Education group at our children’s hospital developed a video about feeding tube placement as an additional resource for families. The 25-minute video describes the experiences of 5 families with feeding tubes. Families with children of different ages, diagnoses, and ethnic backgrounds were interviewed in their homes about their experiences making the decision and living with a g-tube. The interview guide was developed by using the results of 2 qualitative systematic reviews.^4,15  This video, freely available with additional written educational content, is on 1 of the most visited pages of the Web site About Kids Health (https://www.aboutkidshealth.ca/Article?contentid=2822&language=English), with >100 000 hits since 2017. “About Kids Health” provides free health education on a variety of pediatric topics and receives >20 million hits per year.

The project team included the institution’s primary g-tube nurse practitioner; 3 pediatricians with clinical and research experience in the care of children with g-tubes, medical decision-making, and complex care; and 4 research assistants.

We used the traditional 16-statement form DCS,¹⁶  including the recommended nonscored question asking about the caregiver’s current choice (https://decisionaid.ohri.ca/docs/develop/Tools/DCS_English.pdf). The 16 statements ask for a response on a 5-point Likert scale; results are averaged and scaled for a total decisional conflict score out of 100, with higher scores reflecting greater uncertainty. According to the scoring manual,¹⁶  total DCS scores <25 are typically associated with implementation of the decision and scores >37.5 are associated with delay. The DCS includes 5 subscales: uncertainty, values clarity, support, how informed the individual feels, and effective decision-making. Subscales are also averaged and adjusted for a maximum subscale score of 100.

We paired the DCS with a short questionnaire asking for demographic information (Supplemental Information). The first 6 families were also asked to provide feedback on burden of completing the DCS and demographics forms.

Eligible caregivers were identified from the g-tube team consult lists and received a “Dear parent” letter introducing the surveys before their g-tube consultation. Willing participants completed the surveys in person on paper or on a study iPad at 2 time points, as described below. iPad surveys were completed within Research Electronic Data Capture (REDCap); data from paper surveys were transcribed into Research Electronic Data Capture. We recruited families until we reached our minimum sample size of 16 families who completed the postconsult and video DCS.

We trialed several measurement-timing strategies to assess caregiver rates of completion and burden of administration, with the video acting as a stand-in for a postconsultation decision aid. Most families were asked to complete the DCS twice on the same day: before the consultation and after the consultation or after the consultation plus video. The study evolved over several phases because of challenges with clinic flow and poor rates of completion of the second survey. In phase 1, families were asked to complete the DCS before and after consultation. In phase 2, families were asked to complete the DCS once after the consultation only. In phase 3, families were asked to complete the DCS after consultation and again after watching the video. In phase 4, families were asked to complete the DCS once before the consultation and again after watching the video.

We described demographic and basic clinical details about the families. We evaluated the difference in preconsult DCS scores by t test between subgroups identified in previous studies about pediatric g-tube placement: children with and without neurologic impairment¹⁷  and children older and younger than 18 months¹⁸  (data recorded in years, so we included children <2 in the younger group). We described scale completion rates on the basis of timings of administration: before consultation, after consultation, after video. We reported the number of caregivers at each administration time point with scores <25, which are associated with decision implementation, and >37.5, which are associated with delay. We also described DCS scores and decision aid target subscales (informed and values clarity) at each postconsultation time point. Significance was set at the conventional 0.05 level (2 tailed) for bivariate analyses. All analyses were performed in R version 3.6.2.

This project was discussed with our institution’s quality management team. Because the primary goal of the project was to evaluate the decisional support for families during the g-tube consultation process, it met local guidelines for approval through the review of quality improvement projects process and was considered institutional review board exempt.

We enrolled 62 caregivers in the project; however, 1 family chose not to participate after enrollment and was excluded from further analysis. The clinical characteristics of our cohort are described in Table 1. Children had a mean age of 2.25 years (range 1–5), and 45 (73.8%) had neurologic impairment. Community pediatricians were the most common referring physicians (19.4%), and “losing or not gaining weight” was the most common reason for referral (55.7%).

We completed 4 plan-do-study-act cycles testing different administration strategies. Figure 1 reveals the number of families enrolled in each phase and the number of caregivers who completed the scale. All caregivers completed the first administration of the DCS, with 39 completing it before consultation and 23 completing it after the consult. For the second administration, uptake was lower. Most (5 of 6 caregivers [83%]) completed the second DCS when it was administered immediately after the consult. When the first DCS was administered after the consult and the second after the video, no families completed the second DCS (0 of 9). When the first DCS was administered preconsult and the second after the video, approximately half of the caregivers (17 of 33 [52%]) completed the second DCS.

Phase 1 families (n = 6) reported that the demographics questionnaire plus DCS was acceptable (90% found it easy to understand and 100% thought it was the right length), although some families noted that the scale prompts were difficult to answer preconsult. Families were not asked whether the second DCS administration was burdensome, but many caregivers opted not to complete it again.

The mean preconsult DCS total score across all phases for 37 caregivers was 38.7 (SD = 19.4; range 4.7–84). Seven caregivers (19%) had a preconsult DCS score <25; 19 caregivers (51%) had a score >37.5. Of the 5 DCS subscales, caregivers scored highest on the uncertainty subscale (mean = 45.0; SD = 24.6). The informed and values clarity subscale mean scores were 44.1 (SD = 23.7) and 41.4 (SD = 22.8), respectively. The preconsult DCS scores of clinically relevant subgroups are compared in Table 2. There were no significant differences in total baseline scores between caregivers of children with and without neurologic impairment and of children older and younger than 2 years.

Table 3 reveals the DCS scores after consultation and after consultation plus video. The mean total scores were 17.9 (SD 13.5) and 12.7 (SD 13.2) for postconsult and postconsult plus video, respectively. Subscale scores for uncertainty and informed and values clarity were <25, except for uncertainty postconsult, which was 29.8 (SD 21.1). After consultation with or without video, only 3 caregivers (7.7%) had total DCS scores >37.5 associated with decisional delay.

Decisional conflict among caregivers deciding about g-tube placement was measurable, and there were high caregiver completion rates for the first administration in the encounter. However, we were unable to find a practical strategy to facilitate pre–/post–in-person intervention testing because many families chose not to complete the second scale, potentially because of the length of the encounter. Future studies using the DCS should be designed to require only 1 administration or, if 2 administrations are necessary, should consider alternative strategies to increase uptake of the second survey, including incentives, spaced administration over 2 visits, or remote administration.

The mean preconsult score on the DCS (38.7; SD = 19.4) was above threshold for “feeling unsure about implementation” (37.5),¹⁶  which is not unexpected considering it was measured prediscussion of g-tube risks and benefits. However, the range in DCS scores was wide: at initial g-tube consult, some families had low decisional conflict (total scores <10) and other families had high decisional conflict (total scores ≥75). We could not explain this variability with subgroup evaluation: baseline scores were not different among caregivers of children with neurologic impairment or older children. However, that lack of difference between subgroup scores was likely secondary to small sample sizes. The wide variation suggests that the decision-making process for some families may have begun before presentation at g-tube clinics, for example, in conversations with their primary care provider when the referral was made. Additionally, after our institution’s current decision support consultation process, few families (7.7%) had high residual decision conflict (total scores >37.5). Better understanding these families with high residual conflict may provide insights into how to better support all families. Additionally, decisional conflict is only a single dimension of decision-making, and further work is needed to understand what type of decisional support would be helpful and the best timing of its implementation.

The challenges for families in decision-making about g-tube feeding are well described,¹⁹  particularly for children with developmental disabilities.^4,20  However, in no previous studies have researchers evaluated decisional conflict before and after g-tube consultation. In 1 previous study (a doctoral thesis from 2003), the author measured decisional conflict among mothers at the time of g-tube placement at our institution.²¹  In that study, the author reported a mean (SD) score for the uncertainty subscale (updated by using current scoring methods) of 42.5 (25), which is similar to our findings of high preconsult uncertainty subscale score and wide variance (mean = 45.0; SD = 24.6). However, that study’s recruitment occurred at the time of g-tube placement, which suggested families had lingering postdecision uncertainty. Although it is tempting to attribute the much lower post–consult plus video DCS scores in this study (mean 12.7; SD = 13.2) to improvements in the g-tube consultation process, how decisional conflict changes between consultation and g-tube placement is unknown and warrants further evaluation.

This project’s findings have important limitations. First, this effort was framed as a quality improvement project, so our outcomes are institution specific and may not be generalizable. Second, our project was inadequately powered to draw conclusions from the finding of no subgroup score variability. Our goal with this report is to provide feasibility recommendations and baseline estimates for sample size calculations for future studies.

This project reveals that measuring decisional conflict among families considering g-tube placement is feasible at least once during an encounter. There is a wide range in baseline decisional conflict scores, which suggests that further work could be used to explore identification of subgroups who might benefit from targeted decision support. Finally, after completion of our institution’s current decision support consultation process, a few families still have high residual decisional conflict, suggesting that a subset may benefit from additional support. Further work used to characterize decisional conflict and other dimensions of decision-making in this population is necessary to inform development of decision support interventions.

We thank Yuna Lee, MA, Toronto District School Board; Carol Chan, BSc (Hons), Child Health Evaluative Sciences, SickKids Research Institute; and Mitchell Cooper, MSc, Schulich School of Medicine and Dentistry, Western University of Ontario, for their help with data collection and analysis. We also appreciate editing support from Nancy Nelson, MSEd.