Associations of Patient and Parent Characteristics With Parental Decision Regret in the PICU: A Secondary Analysis of the 2015–2017 Navigate Randomized Comparative Trial

Abstract

OBJECTIVES:

To identify self-reported meaningful decisions made by parents in the PICU and to determine patient and parent characteristics associated with the development of parental decision regret, a measurable, self-reported outcome associated with psychologic morbidity.

DESIGN:

Secondary analysis of the Navigate randomized comparative trial (NCT02333396).

SETTING:

Two tertiary, academic PICUs.

PATIENTS:

Spanish- or English-speaking parents of PICU patients aged less than 18 years who were expected to remain in the PICU for greater than 24 hours from time of enrollment or who had a risk of mortality greater than 4% based on Pediatric Index of Mortality 2 score.

INTERVENTIONS:

None.

MEASUREMENTS AND MAIN RESULTS:

Between April 2015 and March 2017, 233 parents of 209 patients completed a survey 3–5 weeks post-PICU discharge which included the Decision Regret Scale (DRS), a 5-item, 5-point Likert scale tool scored from 0 (no regret) to 100 (maximum regret). Two hundred nine patient/parent dyads were analyzed. The decisions parents reported as most important were categorized as: procedure, respiratory support, medical management, parent–staff interactions and communication, symptom management, fluid/electrolytes/nutrition, and no decision. Fifty-one percent of parents had some decision regret (DRS > 0) with 19% scoring in the moderate–severe range (DRS 26–100). The mean DRS score was 12.7 (SD 18.1). Multivariable analysis showed that parental Hispanic ethnicity was associated with greater odds ratio (OR 3.12 [95% CI, 1.36–7.13]; p = 0.007) of mild regret. Being parents of a patient with an increased PICU length of stay (LOS) or underlying respiratory disease was associated with greater odds of moderate–severe regret (OR 1.03 [95% CI, 1.009–1.049]; p = 0.004 and OR 2.91 [95% CI, 1.22–6.94]; p = 0.02, respectively).

CONCLUSIONS:

Decision regret was experienced by half of PICU parents in the 2015–2017 Navigate study. The characteristics associated with decision regret (parental ethnicity, PICU LOS, and respiratory disease) are easily identifiable. Further study is needed to understand what contributes to regret in this population and what interventions could provide support and minimize the development of regret.

Parents of children in the PICU face a variety of medical and nonmedical decisions (1). Decisions, their outcomes, and circumstances surrounding them can lead to decision regret—“the negative, cognitively-based emotion that we experience when realizing or imagining that our present situation would have been better had we acted differently” (2). Decision regret may affect psychologic outcomes like anxiety, depression, and posttraumatic stress disorder (2, 3). A 2016 systematic review found 59 studies using the validated Decision Regret Scale (DRS) to measure regret related to a specific, nonhy-pothetical medical decision (3). Of these 59 studies, most involved adult patients with cancer; only one was conducted in intensive care and one in pediatrics (4, 5). Since this review, there have been additional studies of decision regret in pediatrics, but only two included PICU patients and both focused on specific decisions (tracheostomy and genomic testing) (6, 7). Thus, there is a paucity of data about parental decision regret related to decision-making for PICU patients.

We have previously described decision regret using the DRS as part of the 2015–2017 Navigate randomized comparative trial (NCT02333396) (8). In that study, we reported the average decision regret score for each of the study arms. This does not provide an in-depth understanding of parental decisions or other factors potentially impacting regret. The goal of this post hoc secondary analysis was to examine decision regret in parents of PICU patients by identifying decision regret in a curated dataset from the Navigate study. We also examined parent-reported meaningful decisions and explored patient and parent characteristics associated with decision regret.

MATERIALS AND METHODS

This is a post hoc secondary analysis, completed in 2021, of the 2015–2017 Navigate Study data (registered at ClinicalTrials.gov, ID NCT02333396) (8). This analysis, prompted by the study team’s interest in using the data obtained in the Navigate Study to further explore decision regret, began after publication of the original analysis in 2020. The portion of the analysis assessing types of decisions reported by parents was completed simultaneously with the primary study analysis. The Navigate Study was a two-site, randomized, comparative trial comparing outcomes for parents who received an intervention, PICU Supports, with those who received an informational brochure (8). PICU Supports assisted parental communication, decision-making, emotional, informational, and discharge or end-of-life care needs by using a trained navigator. The Lurie Children’s Hospital Institutional Review Board (IRB) approved “The Navigate Study” and all subsequent analyses (IRB 2015-146) on November 14, 2014. The University of Chicago Biological Sciences Division IRB approved “Improving Communication in the PICU for Patients Facing Life-Changing Decisions: The Navigate Study” (IRB 15-0624) on September 24, 2015. The Washington University in St. Louis IRB approved “Decision Making and Decision Regret in the PICU” (IRB 201712011) on December 22, 2017. Some of the secondary data analysis presented here was carried out at this institution. All study procedures were completed in accordance with the ethical standards of the responsible committee on human experimentation and with the Helsinki Declaration of 1975.

The Navigate Study took place at two university-based, tertiary-care children’s hospitals. The study included Spanish- or English-speaking parents of PICU patients younger than 18 years expected to remain in the PICU for greater than 24 hours from the time of enrollment or who had a risk of mortality greater than 4% based on Pediatric Index of Mortality 2 (PIM2) scoring. The publication of the original study results describes the exclusion criteria and enrollment details (8). Parents provided written consent and patients 12–17 years old provided written assent unless incapable (determined by PICU attending).

Parents completed questionnaires at the time of enrollment and 3–5 weeks after PICU discharge. The enrollment questionnaire included demographic questions and Patient Reported Outcomes Measure Information System (PROMIS) scales for informational social support, depression, and anxiety (9). The 3–5 weeks questionnaire included the PROMIS scales and the DRS. The DRS is a 5-item, 5-point Likert scale tool in which scores are transformed into a 0 (no regret) to 100 (maximum regret) scale. Three of the items are phrased in the positive direction and two in the negative direction to avoid bias. Parents were asked to “write down the MOST important decision made for your child while he/she was in the PICU” and complete the DRS in reference to this decision. Details of scoring, reliability, and validity for all scales were described previously (8). Patient data obtained from chart review included age, underlying conditions, PIM2 score, reason for PICU admission, PICU length of stay (LOS), and presence of any of the following serious PICU events: tracheostomy or gastrostomy tube placement; initiation of chronic ventilation; new do not resuscitate order; initiation of therapeutic limitations; new palliative care consult; and use of cardiopulmonary resuscitation, extracorporeal membrane oxygenation, or continuous renal replacement therapy. Patient outcome was divided into four categories: death, serious PICU event, PICU LOS less than 12 days without a serious event, or PICU LOS greater than or equal to 12 days without a serious event.

We analyzed data from patient/parent dyads if parents completed both the enrollment questionnaires and the DRS at 3–5 weeks post-PICU discharge. Parents did not have to name a most important decision to be included in the analysis. We included these parents because many of them had regret; we reasoned that naming a decision could be hard and there may be value in further understanding their regret. Two authors separately reviewed the decisions named by parents and created categories. They compared categories, combining redundant categories as needed. Each decision was then assigned to a category by the two reviewers. Discrepancies were resolved by consensus. Validated questionnaires were scored per instrument instructions. We divided the DRS scores into no regret (DRS 0), mild regret (DRS 1–25), and moderate–severe regret (DRS 26–100), based on a cutoff of 25 (between mild and severe regret). This strategy has been used and subsequently validated by others (3, 10).

Statistical analysis was performed with SAS Enterprise Guide 7.15 software (SAS Institute, Cary, NC). Chi-square or Fischer’s exact tests were used for categorical variables and analysis of variance for continuous variables. Bivariate analysis was used to determine associations between patient and parent characteristics and the primary outcome, decision regret, which had three outcome levels (DRS = 0, DRS = 1–25, or DRS = 26–100). For parent characteristics, we used results from one randomly selected parent per patient to ensure independence between surveys. Variables with a p value of less than 0.1 in bivariate analysis were included in a multinomial logistic regression model to explore associations with the three different levels of regret. Included variables were chosen in this way because prior pediatric studies have either not investigated demographics in relation to regret or have found inconsistent results, making an informed a priori selection of variables difficult. Alpha level for significance was 0.05. DRS of 0 (no regret) was used as the reference category.

RESULTS

In the Navigate study (April 6, 2015 to March 7, 2017) 4251 patient families were screened. Postdischarge surveys were received from 241 parents of 216 patients (56.5% of families enrolled). After randomizing one parent for each patient and excluding those who did not complete the DRS, we analyzed complete data from 209 patients and parents (Fig. 1). Table 1 describes the demographics and clinical characteristics of the patients. Parent characteristics can be found in Table S1S (http://links.lww.com/PCC/C517). Median parent age was 41 years with 82% (171/209) female parents and 62% (130/209) of parents identifying as White, 18% (38/209) as Black, and 18% (38/209) as Hispanic.

Figure 1.

Study enrollment. DCFS = Department of Child and Family Service.

TABLE 1.

Patient Demographics and Clinical Information

Characteristics	Overall (n = 209)	DRS = 0 (n = 102)	DRS: 1–25 (n = 67)	DRS: 26–100 (n = 40)
Age at admission (yr), median (IQR)	2.2 (0.5–8.8)	1.9 (0.4–8.2)	2.7 (0.6–11.3)	1.8 (0.6–7.4)
Pediatric Index of Mortality 2 score, median (IQR)	1.2 (0.7–3.9)	1.2 (0.6–3.7)	1.1 (0.6–3.2)	1.5 (0.9–4.6)
Length of stay (d), median (IQR)	9 (6–17)	8 (6–13)	10 (5–17)	14 (8–30)
Number of times hospitalized, median (IQR)	2 (0–5)	1 (0–5)	2 (0–7)	2 (1–5)
Number of times in PICU, median (IQR)	1 (0–3)	0 (0–2)	1 (0–3)	2 (0–3)
Presence of underlying diseasea, n (%)
Neuromuscular	71 (34.0)	28 (27.5)	25 (37.3)	18 (45.0)
Cardiovascular	83 (39.7)	41 (40.2)	22 (32.8)	20 (50.0)
Respiratoryb	62 (29.7)	22 (21.6)	24 (35.8)	16 (40.0)
Renal	18 (8.6)	9 (8.9)	4 (6.0)	5 (18.5)
Gastrointestinal	68 (32.5)	30 (29.4)	22 (32.8)	16 (40)
Hematology/immunodeficiency	8 (3.8)	3 (2.9)	2 (3.0)	3 (7.5)
Metabolic	16 (7.7)	6 (5.9)	6 (9.0)	4 (10.0)
Endocrine	14 (6.7)	5 (4.9)	6 (9.0)	3 (7.5)
Rheumatologic	3 (1.4)	2 (2.0)	0 (0)	1 (2.5)
Other congenital	31 (14.8)	15 (14.7)	12 (17.9)	4 (10.0)
Malignancy	14 (6.7)	5 (4.9)	5 (7.5)	4 (10.0)
Stem cell transplant	4 (1.9)	0 (0)	3 (4.5)	1 (2.5)
Otherwise healthy	33 (15.8)	20 (19.6)	10 (14.9)	3 (7.5)
Reason for admissiona, n (%)
Acute respiratory illness	75 (35.9)	31 (30.4)	30 (44.8)	14 (35.0)
Acute septic illness	11 (5.3)	4 (3.9)	3 (4.5)	4 (10.0)
Acute neurologic illness (not trauma)	13 (6.2)	9 (8.8)	2 (3.0)	2 (5.0)
Elective admit	2 (1.0)	2 (2.0)	0 (0)	0 (0)
Fluid/electrolyte issues	14 (6.7)	6 (5.9)	5 (7.5)	3 (7.5)
Ingestion	1 (0.5)	1 (1.0)	0 (0)	0 (0)
Other	27 (12.9)	15 (14.7)	7 (10.4)	5 (12.5)
Trauma	4 (1.9)	1 (1.0)	1 (1.5)	2 (5.0)
Postoperative	68 (32.5)	35 (34.3)	21 (31.3)	12 (30.0)
Parent perception of illness severityc, n (%)
Mild	10 (4.8)	2 (2.0)	4 (6.0)	4 (10.0)
Moderate	75 (35.9)	41 (40.0)	25 (37.3)	9 (22.5)
Severe	114 (54.5)	57 (55.9)	33 (49.3)	24 (60.0)
Missing response	10 (4.8)	2 (2.0)	5 (7.5)	3 (7.5)
Patient outcome, n (%)
Patient died	5 (2.4)	0 (0)	0 (0)	5 (12.5)
Patient experienced serious event	9 (4.3)	5 (4.9)	2 (3.0)	2 (5.0)
Patient without serious event and length of stay ≥ 12 d	81 (38.8)	32 (31.4)	31 (46.3)	18 (40.0)
Patient without serious event and length of stay < 12 d	114 (54.5)	65 (63.7)	34 (50.7)	15 (37.5)

DRS = Decision Regret Scale, IQR = interquartile range.

^aPatient could fall in multiple categories, therefore percentage for each category is presented instead of for the variable overall.

^bIncludes respiratory malformations, chronic respiratory disease, cystic fibrosis, tracheomalacia, pulmonary hypertension, bronchopulmonary dysplasia, obstructive sleep apnea, and trach/vent dependence.

^cOne hundred ninety-nine of 209 parents completed this survey question.

Parent-reported decisions fell into seven categories: procedures, respiratory support, medical management, parent–staff interactions and communication, symptom management, fluid/electrolytes/nutrition, and no decision (Table 2). The most named decisions involved procedures, respiratory support, and medical management. The average decision regret score for all decisions was 12.7 (sd 18.1) with a median of 5 (interquartile range 0–20). 49% (102/209, [95% CI, 42–56%]) of parents had no regret, 32% (67/209, [95% CI, 26–38%]) had mild regret, and 19% (40/209, [95% CI, 14–24%]) had moderate–severe regret.

TABLE 2.

Parent-Reported “Most Important” Decisions

Decision Category	Subcategory Examples	Examples of Parent-Named Decisions	Number (% of Parentsa)	Decision Regret Score (Median, Interquartile range)
Procedure	Gastrostomy tube, tracheostomy, surgical procedure, peripherally inserted central catheter placement	“Receiving the stoma revision surgery” “Draining fluids from his chest”	68 (29.2%)	0 (0–15)
Respiratory support	Bipap initiation, intubation, extubation	“When to extubate” “Whether or not to re-intubate after failed extubation”	67 (28.8%)	5 (0–25)
Medical management	Medication choice, seizure management, treatment timing	“To put her on a daily dose of catopril when she was discharged” “Whether to start blood thinners”	57 (24.5%)	0 (0–19.4)
Parent–staff interactions and communication	Sharing patient information with team, discharge planning, parent involvement in care	“To communicate to the healthcare team how to care for my child because she is sensitive. I told the healthcare team I wanted to give her the medications myself.” “To not send him home until the staff of doctors and nurses were 100% sure he was ready and had the oxygen level he needed and all their T’s crossed and I’s dotted”	20 (8.6%)	0 (0–27.5)
Symptom management	Patient comfort, pain management	“How to keep my child the most comfortable while waiting for transplant” “Keeping his pain under control”	15 (6.4%)	0 (0–25)
Fluid, electrolytes, nutrition	Feeding type or timing, fluid management	“Giving her time to eat/drink on her own (no NG tube)” “Changing her formula”	14 (6.0%)	2.5 (0–10)
No decision	Did not name a decision, wrote not applicable, or answer did not describe a decision	“Everything was important for me” “Cardiac surgeons were fantastic”	43 (20.2%)	15 (0–30)

^aSome parent answers described more than one decision and so were placed in more than one category.

Bivariate analysis showed several patient and parent characteristics associated with the presence of decision regret: PICU LOS, presence of an underlying respiratory disease at time of admission, parental Hispanic ethnicity, and patient outcome. Multivariable analysis of variables with a p value of less than 0.1 in the bivariate analysis (ethnicity, informational social support, depression, PICU LOS, patient underlying respiratory disease, patient underlying stem cell transplant, and patient outcome) showed three characteristics that were associated with greater odds of regret (Table 3). Parental Hispanic ethnicity was associated with greater odds of mild regret while PICU LOS and underlying respiratory disease were associated with greater odds of moderate–severe regret.

TABLE 3.

Characteristics Associated With Decision Regret—Multivariable Logistic Regression

Variable	Mild Regret		Moderate–Severe Regret
	OR (95% CI)	p	OR (95% CI)	p
Parental depression score	1.03 (0.99–1.08)	0.16	1.04 (0.99–1.10)	0.15
PICU length of stay	1.00 (0.98–1.03)	0.98	1.03 (1.01–1.05)	0.004
Patient underlying respiratory disease	2.02 (0.97–4.24)	0.06	2.91 (1.22–6.94)	0.02
Patient underlying stem cell transplanta	> 999 (<0.001–> 999)	0.98	> 999 (< 0.001–> 999)	0.98
Patient outcome	1.58 (0.88–2.84)	0.13	0.79 (0.45–1.38)	0.41
Parental informational social support score	0.97 (0.93–1.02)	0.21	0.98 (0.93–1.03)	0.46
Parental ethnicity	3.12 (1.36–7.13)	0.007	1.49 (0.48–4.56)	0.49

OR = odds ratio.

^aNo parent of a patient with underlying stem cell transplant scored in the reference category of Decision Regret Scale = 0. Boldface values indicate statistically significant.

DISCUSSION

In this post hoc secondary analysis of the 2015–2017 Navigate Study comparative trial dataset (8), we found that decision regret affects approximately half of parents of patients admitted to the PICU. Our analyses also identified parent-reported decisions and explored the patient and parent characteristics associated with regret. The decisions reported were most frequently related to procedures, respiratory support, or medical management. In multivariable analysis, PICU LOS, underlying respiratory disease, and parental Hispanic ethnicity were associated with greater odds of decision regret. Although these factors may not be modifiable, they are identifiable, allowing future work to investigate contributors to regret in these at-risk populations and to develop targeted interventions aimed at decreasing the prevalence of regret.

In this study, the mean regret score of 12.7, was consistent with observations in a systematic review of literature through 2014, which included primarily adult patient studies (3). Across these studies, the mean DRS was 16.5 with individual study means ranging from 2.5 to 49.1. In a subset of these studies that compared a DRS of greater than 0 to a DRS of 0, a median of 37% of participants experienced regret. A review of studies up to 2018, assessing decision regret in caregivers of adult and pediatric patients found a median mean DRS of 14.3 across the studies (11). The prevalence of regret in our study was 51%. Two other studies of decision regret in parents of critically ill children, one reflecting decisions about tracheostomy and one regarding consent for genomic testing, showed a prevalence of 52% and 42%, respectively (6, 7). Despite a similar mean score to other studies, the percentage of parents in our population who experienced regret is notable. In a subset of adult studies that compared DRS greater than 25 to DRS less than or equal to 25, the percentage of participants experiencing moderate–severe regret ranged from 4% to 20% (3). In a study of parents of critically ill children with neurologic conditions, 20% of parents had moderate–severe regret, similar to the 19% in our study (12). As our analysis used data from 2015 to 2017, future work should verify this observation in the contemporary era as well as investigate potential contributory factors, such as the decisions made in the PICU and the complexities related to making decisions as a surrogate for a child. This study focused on identifying parents with regret, but also important are the 49% of parents who experienced no regret. Interestingly, a recent study of bereaved parents revealed that five of six parents did not regret ventricular assist device implantation despite their child’s death (13). Future studies should explore resilience, decision-making, and nondevelopment of regret among parents who do not report regret.

Our analysis showed greater odds of decision regret associated with parental Hispanic ethnicity, PICU LOS, and underlying respiratory disease. Several recent studies have focused on associations between decision regret and components of the decision-making process such as decisional conflict, satisfaction with information, preferred role in decision-making, or perceived quality of communication (3, 12-17). Other studies have looked at regret in relation to decision outcomes (18-20). The results of these studies are inconsistent across populations and decisions. Regarding sociodemographic variables, a study of parents of pediatric oncology patients also demonstrated that Hispanic ethnicity was positively correlated with regret. This same study showed a positive correlation between regret and parents who self-identified as Black, which we did not find in our analysis (21). We found no other pediatric studies that assessed the association of PICU (or hospital) LOS or chronic comorbidities with regret (15, 21).

Since many studies show that prevalence of regret is low and most regret is mild, identifying demographic variables associated with parental decision regret facilitates focused study of process measures in at-risk groups. These demographic variables could direct further study and, subsequently, implementation of additional supports for parents during and after the decision-making process. Development of these supports would require additional work using methods such as interviews and focus groups to identify any consistent factors contributing to regret in these groups and to understand, from parental perspectives, what supports would be most impactful.

One unique aspect of the Navigate Study was that parents named the decision most important to them. Soltys et al (22) used a similar approach in parents of former neonatal intensive care (NICU) patients. In that study, many of the decision categories identified were similar to those we describe (e.g., decisions related to medications, procedures, or feeding). The most commonly identified category by Soltys et al was the nonmedical decision related to separation of the parent from their infant (19.3% of decisions). The category of nonmedical decisions identified by parents in our study, parent–staff interactions and communication, was reported by only 8.6% of parents. One of the most common decision categories described by parents in our study was related to respiratory support (28.8%), a category not identified by Soltys et al. Interestingly, in both studies nearly 20% of parents did not name a decision. In our study, these parents still completed the DRS, potentially suggesting that the concept of feeling positively or negatively about decisions resonated with them, even if they did not name a specific decision. The missing decision could be explained by parents reflecting on general feelings of regret related to the ICU experience rather than a specific decision, not being aware of what decisions were made, not being able to narrow down to one decision, or feeling so negatively about a decision that they could not write it down. Because these parents also reported the highest median regret score, future work should address this population using more open-ended strategies such as interviews.

Our findings that parents identifying as Hispanic and parents of children who have longer PICU stays or underlying respiratory disease had greater odds of reporting higher decision regret warrants further comment. Given that a decreased LOS is viewed as a marker of a good outcome in many PICU studies, it makes sense that an increased LOS, as a surrogate marker of poor outcome, might result in decision regret. Certainly, increased LOS is often associated with higher acuity patients, and parents of high acuity patients may face more difficult and impactful decisions. A longer PICU LOS may increase the possibility that parents face more and/or more complex decisions. Patients with underlying respiratory diseases may also demonstrate more medical complexity, potentially exposing parents to more complex decisions. Regarding parents identifying as Hispanic, we did not ask about language fluency or preferred language. Increased regret in this population could have resulted from language barriers impacting communication and should be assessed in future studies.

In addition to the previously described limitations of the Navigate Study (8, 23), we acknowledge limitations of this secondary analysis. Having parents identify the most important decision, rather than selecting a decision for them, provided valuable information. However, analyzing data from different decision types may have diluted our ability to understand the influences on and factors associated with decision regret. We chose this approach as it reflects the experience of the general PICU population and may have broader applicability. Our approach also required parental identification of a decision, which 20% of parents did not do. There is no standard time to assess regret. We assessed decision regret 3–5 weeks following PICU discharge, but in the systematic review discussed above (3), regret was assessed anywhere from immediately following the decision to 9 years later, with some studies showing increasing regret over time. Also, determining the meaning of regret is complex. Although regret is generally considered a negative emotion and is associated with psychologic morbidity, some authors argue that regret has benefits (24). For parents of PICU patients, regret could motivate them to approach decision-making differently during a subsequent PICU admission, something relevant for parents of children with medical complexity who may comprise up to 77% of the PICU population (25). Finally, this study does not explore those 49% of parents with no regret, an important population that could help identify factors that protect against regret.

CONCLUSIONS

In the 2015–2017 Navigate, comparative PICU study cohort, parents reported a variety of decisions as being the most important, and approximately half of these parents reported decision regret. One in five (95% CI, 14% to 24%) experienced high levels of regret. Parental Hispanic ethnicity, patient PICU LOS, and patient underlying respiratory disease were each associated with greater odds of parental decision regret. These findings demonstrate the importance of further in-depth study of PICU parents to better understand key drivers of decision regret. Furthermore, such studies may serve as a foundation to develop interventions that support at-risk groups and minimize the development of decision regret and its associated psychologic morbidities.

Supplementary Material

Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s website (http://journals.lww.com/pccmjournal).

RESEARCH IN CONTEXT.

• Measures of decision regret are increasingly being used as a patient-centered outcome, particularly in adult oncology.

• Research on decision regret in surrogate decisionmakers such as parents is less common. Research in pediatrics is limited to small studies focused on a specific population or decision.

• This study aimed to assess decisions made by and decision regret in PICU parents in the Navigate trial dataset from 2015 to 2017.

AT THE BEDSIDE.

• PICU parents reported a broad spectrum of decisions that they considered the most important decision they made for their child while in the PICU.

• Half of parents experienced decision regret. As regret has been shown to affect experience with the healthcare system, this may be relevant to our interactions with families and is something we should consider when supporting them in the PICU.

• Hispanic ethnicity, PICU length of stay, and patient underlying respiratory disease were associated with greater odds of decision regret. Identifying at-risk groups could direct implementation of additional supports.