An assessment of clinical and system drivers of family satisfaction in the pediatric intensive care unit

Abstract

Objective:

Investigate clinical and system drivers of family satisfaction in the PICU.

Design:

Mixed methods qualitative and quantitative (observational) study. Qualitative interviews with families were performed as a pilot to inform modality of survey distribution based on family preferences. A validated pediatric satisfaction survey deployed to family members for seven months with a corresponding chart review and administrative data collection.

Setting:

Pediatric intensive care unit in a tertiary children’s hospital

Patients:

206 families of patients admitted to the PICU >48 hours.

Interventions:

None.

Measurements and Main Results:

Families preferred surveys distributed electronically on a tablet in the PICU setting. The Pediatric Family Satisfaction-ICU survey was used to assess comfort with medical decision-making and communication with the care team. Capture rate of all eligible patients was 69.5% and response rate of 90.8%. Overall, 64.7% of respondents were highly satisfied, while over one third were not highly satisfied; families of Hispanic ethnicity (odds ratio of lower satisfaction of families with Hispanic ethnicity: 2.09, 95% CI: 1.01, 4.33; p= 0.047) and high social stressors (odds ratio of higher satisfaction among high stressed subgroup: 0.49, 95% CI: 0.24, 0.99; p=0.047) reported statistically significant lower satisfaction. Additional free text responses were identified in 21% of respondents, with the majority of comments indicating wishes for improvements clustered around communication with the medical team or sleeping environment of families and patients.

Conclusions:

High capture rates of family satisfaction in the PICU can be obtained with a PICU-specific survey, limiting barriers to completion by including family preferences, and distributing in the PICU setting. Less than 2/3 of PICU families are highly satisfied; patients of Hispanic ethnicity and those with high social stressors predict low satisfaction, while illness severity, age and PICU LOS did not have statistical significance. Local improvement teams can use this approach to drive enhanced satisfaction.

Introduction

Over 150,000 children are admitted to pediatric intensive care units (PICU) annually in the United States at a substantial emotional and financial burden for families and children’s hospitals (1–4). Over the past three decades, pediatric critical care mortality has decreased by nearly 50% (5). As a result, research and discovery efforts have shifted from being primarily mortality-focused to assessing effectiveness of healthcare delivery (6). Family engagement initiatives are increasing and quality improvement (QI) and research collaboratives are now integrating family input, including efforts towards liberation from the harmful effects of an ICU stay (7).

To assess these efforts, patient and family satisfaction has become an area of increased focus for healthcare systems (8–11). Understanding satisfaction data improves healthcare teams to provide compassionate care as families face challenging decisions and insures sharing of family goals during the trajectory of their family child’s course (12). Insight into family satisfaction increases awareness of care expectations are and improves sensitivity to patient characteristics and situations that may indicate poor perception of care (11, 13). Additionally, measuring patient and family perspectives as an outward facing indicator of healthcare quality has been endorsed by leading professional organizations, including the National Academy of Medicine and Society of Critical Care Medicine, among others (9, 12–13). Studies conducted with adult ICU patients and families indicate higher satisfaction is correlated with lower in-hospital mortality, lower readmission rates, and follow-up with specialists upon discharge (14–16). While this signals potential use of satisfaction data as a process metric for downstream patient follow-up or compliance to care plans, patient satisfaction in itself represents an outcome of merit to provide excellent patient care. The National Academy of Medicine has proposed patient satisfaction is “a key component for enhancement of the effectiveness of healthcare microsystems” (12–17).

However, while quality improvement efforts are now standard across PICUs, patient and family satisfaction represents a key health outcome that lacks consensus definition in children, including the determinants of high or low satisfaction (18–19). Most studies assessing PICU satisfaction to this point have primarily aimed to validate tools on small samples, and there remains a gap in recognizing clinical or system factors that correlate with satisfaction (19). Recent studies indicate early success studying ICU satisfaction, but challenges exist (20–22). Much of the challenge lies in measuring satisfaction in the inpatient setting. Patients often spend time on multiple floors or units during a single admission, yet surveys target family or patient-reported satisfaction, often based on the overall hospital encounter, without capturing the specific experience of families in the PICU (23).

The purpose of this study was to investigate clinical and system drivers of family satisfaction in the PICU. Additionally, we aimed to determine an ideal method of satisfaction survey distribution in the PICU. We surveyed caregivers of pediatric patients at the time of transfer from the PICU, in coordination with a focused review of the patient’s hospital course, to detect relationships between populations in the PICU and family satisfaction. We hypothesized populations with older age, lower severity of illness, and shorter PICU length of stay (LOS) would be associated with higher satisfaction.

Materials and Methods

Over seven months, we utilized a multi-pronged approach to gather family satisfaction data in real-time in a 44-bed PICU at a large academic children’s hospital. The process for collecting timely satisfaction data involved a population-specific survey, piloting the survey implementation with families, and distributing the survey in the PICU setting specifically. We then extracted system and clinical variables from patient medical records and Virtual Pediatric Systems ^LLC to link family satisfaction responses with patient data (24). The University of Utah and Intermountain Healthcare Institutional Review Boards reviewed and approved the study and informed consent was obtained prior to survey completion.

Instrument

The Family Satisfaction-ICU was developed for use in adult ICUs and has been since modified and validated in PICUs as the Pediatric Family Satisfaction-ICU (pFS-ICU; Supplemental Digital Content 1) (20–22). Evaluation of the pFS-ICU shows high internal consistency and construct validity for assessing caregiver satisfaction in the PICU (20). The survey consists of 24 Likert-scale questions to assess satisfaction with communication of care and medical decision-making, three open-ended questions regarding opinions, and three respondent demographic questions.

Implementation Pilot

Prior to the final pFS-ICU distribution, we conducted a pilot of the survey implementation process. We sought to evaluate modality preference of survey distribution (electronically on a tablet versus on paper), monitor timing to complete the survey, and to identify barriers in inpatient survey completion. The tablet version was available in REDCap (Research Electronic Data Capture) and paper copies were printed (25). Over the course of three weeks, families were recruited (based on willingness to participate at time of PICU transfer or discharge) to take the survey and participate in a brief in-person interview. Primary caregivers received and completed the survey either on paper or tablet, alternating sequentially between modalities. Participants were qualitatively observed during survey administration to identify barriers to completion, formatting difficulties, and timing to complete. Upon completion, participants were interviewed about mode preference, usability, and comfort with tablet or paper use.

Data Collection: pFS-ICU Distribution

Subjects were family members of patients in the PICU from April through November 2018, both English and Spanish speaking (in-person Spanish interpretation is available during all hours within our institution). Inclusion criteria was patient age 0-18 years, family member completing the survey age ≥18 years, and PICU length of stay ≥48 hours. Exclusions were mortalities, direct admission from the NICU, or prior inclusion during study duration. The study was conducted in the medical PICU, which excludes most children with congenital heart disease. Families were approached to respond to the pFS-ICU in the PICU once the decision had been made to transfer or discharge from the PICU.

Distribution was conducted by research associates from a pediatric research program at the study institution (the associates were not involved in clinical care). The project team, front-line providers, and unit leadership identified an approach for daily survey distribution, involving the research associates obtaining a daily list of discharges from the unit charge nurse. Families were approached as they awaited discharge or transfer from the PICU, during days, nights, and weekends. A single family member completing the survey was deemed adequate for completion. Spanish-translated surveys and in-person interpreters were available for families whose preferred language was Spanish.

Chart Review

Once the primary caregiver completed the pFS-ICU, a detailed manual chart review of the patient was performed and analyzed to assess factors associated with high or low satisfaction. Chart review was supplemented with institutional use of Virtual Pediatric Systems (24). The review included two categories of data, “Patient Factors” and “System Factors”. Patient factors were: demographics, admission diagnosis, severity of illness (PRISM III scoring assessing likelihood of mortality), children with medical complexity, ventilator days, the use of ECMO, vasoactive medications, neuromuscular blockade, CPR, or multiple intubations during PICU stay (26). Children with medical complexity were divided into groups of no complexity or chronicity, chronic but no complexity, or chronic and complex, as defined by Simon, et al, and The Center of Excellence on Quality of Care Measures for Children with Complex Needs consensus definitions (27). System factors were: PICU length of stay (LOS), mode of arrival to the PICU (planned or post-operative versus out of hospital transfer, rapid response, or code event), time of day of PICU admission and discharge (daytime as 7am-7pm, nighttime as 7pm-7am), volume at time of PICU admission and discharge (“low” indicating below the monthly average volume, versus “normal” indicating at or above the monthly average volume), event reports submitted throughout PICU admission, social work acuity level (described on admission by local unit social workers as low, medium, or high, based on their in-person assessment and perceived severity of admission, with high indicating highest concern of social stressors) and involvement of the palliative care team or the child abuse team. Social work assessment is a subjective, but standardized process; a team of two core PICU social workers assessed families for medical severity, family support or tension, and economic and geographic needs.

Statistical Analysis

Our primary outcome of family satisfaction was dichotomized for analysis to identify characteristics of families without high satisfaction. An a priori power analysis produced a total sample size of 206 to detect a difference in satisfied and less satisfied families with 80% power at 0.05 significance level. This calculation was based on studies reporting age to be associated with patient satisfaction in the adult ICU population and a description of PICU patient age by mean (3.5 years) and standard deviation (1.6) (14, 28).

For the pFS-ICU distribution, we summarized the pFS-ICU capture rate (percent of all eligible families that received the survey) and response rate (the percent of families that received the survey that completed it).

Respondent and patient demographics and system characteristics were summarized using means and standard deviations or medians and interquartile ranges (IQRs) for continuous variables and count (%) for categorical variables. Satisfaction was scored for the total instrument and the two subscales (communication with care and decision-making) by averaging individual items for subjects who had a ≥70% item response rate within each subscale (29). The ranges of the total instrument and the two subscale scores were 0-100, where 100 indicated perfect satisfaction. Scores were dichotomized into high (≥80 on the 0-100 scale) versus low/normal satisfaction (<80 on the 0-100 scale), to identify the cohorts of patients who were not highly satisfied as a target of improvement, which facilitates comparisons to prior studies in which overall satisfaction was near 80% (15). Demographics were summarized and compared across high versus low/normal satisfaction using Wilcoxon rank sum tests for continuous variables and chi-squared or Fisher’s exact tests for categorical variables. Univariate logistic regression models compared each demographic variable to satisfaction. We constructed a multivariate model that included the statistically significant variables and additional variables that were clinically relevant (age, sex, PICU LOS, and severity level). Odds ratios (ORs), 95% confidence intervals (CIs) and p-values were reported from the models. Likelihood ratio tests were conducted for categorical variables with more than two levels to approximately check for any significant pair-wise comparisons, including respondent relationship, race and diagnosis.

The two pFS-ICU domains, communication with care and decision-making, were compared using Pearson and Spearman correlations, and 95% CIs were provided. Statistical significance was assessed at the 0.05 level and all tests were two-tailed. Analyses were conducted in R v.3.4.4 (30). Open-ended responses were reviewed by the team to identify themes.

Measurement and Main Results

Implementation Pilot

In the pilot, after completing either the paper or tablet version, 76.5% of interviewed caregivers indicated they would have preferred the tablet version (Supplemental Digital Content 2). Those reporting daily tablet use were significantly more likely to prefer the tablet survey than those without daily tablet use (92.3% vs. 25.0%, p=0.022). Reasons expressed for preferring the tablet over paper included user-friendliness, speed, and perceived enhanced security on tablet. Response time did not differ between paper and tablet. Based on these data, distribution via tablet was deemed appropriate.

pFS-ICU Distribution

Over 200 families (n=206) completed the pFS-ICU (Figure 1). The capture rate was 69.5% of eligible subjects and the response rate was 90.8% of those offered the pFS-ICU. Respondents were predominantly female (72%, N=148) and the primary respondent relationship was the mother (72%, N=147) (Table 1). Patient characteristics show a mean age of 6.9 years (SD 5.8), with ranges from 14 days to 17 years and 11 months. Patients were mostly male (58%, N=119) and white (91%, N=188) (Table 1). Seventeen percent (N=36) of respondents self-identified their ethnicity as Hispanic or Latino; 4% (N=9) requested a Spanish interpreter for completion of the survey. The majority of patients had either complex chronic disease (42%, N=87) or non-complex chronic disease (19%, N=39). Of 7 broad primary diagnosis categories, respiratory was the most common, though neurologic was nearly as prevalent. Co-management with the trauma team (composed of the general pediatric surgery team) occurred for 12% (N=24) of patients.

Figure 1.

Inclusion and exclusion criteria for survey dissemination

Table 1.

Respondent, patient, and system characteristics

Clinical Variables (N=206*)	Respondent Summary	Patient Summary	System Variables	System Summary
Age [mean (SD)]	36.9 (8.9)	6.9 (5.8)	Length of stay in PICU [mean (SD)]	5.1 (6.2)
Sex
Female	148 (72%)	87 (42%)	Arrival time
Male	57 (28%)	119 (58%)	Day	124 (60%)
			Night	82 (40%)
Relationship
Father	46 (22%)
Foster Parent	4 (2%)		Transfer time
Grandparent	4 (2%)		Day	187 (91%)
Mother	147 (72%)		Night	19 (9%)
Other	4 (2%)
			Volume at admission
Ethnicity			High	107 (52%)
Hispanic or Latino		36 (17%)	Low	99 (48%)
Not Hispanic or Latino		170 (83%)
			Volume at transfer
Language			High	97 (47%)
English		197 (96%)	Low	109 (53%)
Spanish		9 (4%)
			Mode of Arrival
Diagnosis Category			Expected	53 (26%)
Respiratory		68 (33%)	Unexpected	153 (74%)
Neurologic		60 (29%)
Other		29 (14%)	SW assessment of high acuity
GI		19 (9%)	Yes	49 (24%)
Infectious		14 (7%)	No	156 (76%)
Oncology		10 (5%)
			Palliative care involvement
Trauma Diagnosis			Yes	38 (18%)
Yes		79 (38%)	No	168 (82%)
No		24 (12%)
			Event Report
PRISM			Yes	65 (32%)
0		79 (38%)	No	141 (68%)
1-3		48 (23%)
4-7		44 (21%)
>7		35 (17%)
Complexity
Non-complex, non-chronic		80 (39%)
Non-complex, chronic disease		39 (19%)
Complex, chronic disease		87 (42%)
Multiple Intubations
Yes		20 (10%)
No		186 (90%)

^*Missing values: Age = 3, Sex = 1, Relationship = 1

Satisfaction Data

Overall, family members of 64.7% of patients were highly satisfied. The two subscales of the survey, the communication with care and decision-making domains were significantly correlated, with a Pearson correlation of: 0.77 (95% CI: 0.71-0.82) with identical Spearman coefficients; this correlation was performed to control for survey fatigue. Individual questions stratified by Likert answer are displayed in Figure 2. Domains assessing inclusion in decision-making, frequency of communication, and support of agitation were among the lowest scores for satisfaction (Figure 2).

Figure 2.

a. Responses per question: Decision-making domain; question number listed from top to bottom in order of increasing percentage of subjects answering “Excellent”. For individual questions in full length, see Supplemental Digital Content 1

b. Responses per question: Communication domain; question number listed from top to bottom in order of increasing percentage of subjects answering “Excellent”. For individual questions in full length, see Supplemental Digital Content 1

Hispanic ethnicity, Spanish-speaking and social work assessment of high acuity were significant for lower satisfaction on univariate analysis. The initial hypotheses for associations with satisfaction, including age (OR: 1, 95% CI: 0.97, 1.03; p=0.86) and LOS (OR: 0.99, 95% CI: 0.95, 1.04; p=0.66), were not statistically correlated with higher or lower satisfaction (Table 2).

Table 2.

Univariate and multivariate analysis comparing all variables with total score satisfaction.

Variables	Univariate Analysis		Multivariate Analysis*
	OR (95% CI)	p-value	OR (95% CI)	p-value
Respondent Age (years)	1 (0.97, 1.03)	0.86	-	-
Respondent Sex – Female	0.81 (0.43, 1.53)	0.52	-	-
Respondent Relationship
Father	Reference	-	-	-
Foster Parent	0.59 (0.08, 4.55)	0.61	-	-
Grandparent	0.59 (0.08, 4.55)	0.61	-	-
Mother	1.17 (0.59, 2.34)	0.65	-	-
Other	0.59 (0.08, 4.55)	0.61	-	-
Patient Age (years)	0.98 (0.93, 1.03)	0.45	0.99 (0.94,1.04)	0.62
Patient Sex – Male	0.85 (0.48, 1.52)	0.59	0.8 (0.43,1.47)	0.48
Race
Caucasian	Reference	-	-	-
Asian	1.48 (0.28, 7.85)	0.64	-	-
Other	5.93 (0.74, 47.31)	0.093	-	-
Ethnicity – Hispanic / Latino	Reference	-	-	-
– Not Hispanic or Latino	2.09 (1.01, 4.33)	0.047	2.22 (1.03,4.78)	0.04
Primary Language – Spanish	0.14 (0.03, 0.71)	0.018	-	-
Diagnosis
Respiratory	Reference	-	-	-
Cardiac	0.51 (0.1, 2.74)	0.43	-	-
GI	1.43 (0.46, 4.47)	0.54	-	-
Infectious	1.28 (0.36, 4.52)	0.70	-	-
Neurologic	0.58 (0.29, 1.19)	0.14	-	-
Oncology	2.04 (0.4, 10.42)	0.39	-	-
Other	1.34 (0.52, 3.49)	0.55	-	-
Trauma Diagnosis – Yes	0.5 (0.21, 1.19)	0.12	-	-
PRISM III Score
0	Reference	-	-	-
1-3	1.66 (0.78, 3.54)	0.19	1.86 (0.86,4.15)	0.12
4-7	1.62 (0.75, 3.51)	0.22	1.66 (0.75,3.79)	0.22
>7	1.89 (0.8, 4.45)	0.15	2.51 (1,6.72)	0.06
Complex
Complex Chronic Disease	Reference	-	-	-
No	0.79 (0.42, 1.47)	0.46	-	-
Non-Complex Chronic Disease	1.85 (0.78, 4.38)	0.16	-	-
Multiple Intubations – Yes	0.81 (0.31, 2.07)	0.65	-	-
Ventilator days	1.02 (0.96, 1.09)	0.44	-	-
PICU length of stay (days)	0.99 (0.95, 1.04)	0.66	0.98 (0.94,1.04)	0.53
Event Report Filed – Yes	0.91 (0.49, 1.68)	0.76	-	-
Time of PICU admission – Night	1.01 (0.56, 1.8)	0.99	-	0.75
Time of PICU discharge/transfer – Night	0.94 (0.35, 2.49)	0.89	-	0.90
Volume of PICU at admission – Normal/Low	1.1 (0.62, 1.94)	0.75	-	0.83
Volume of PICU at discharge/transfer – Normal/Low	1.48 (0.84, 2.63)	0.18	-	0.12
Mode of arrival to PICU – Unexpected	1.27 (0.67, 2.43)	0.46	-	0.32
Social work high acuity	0.5 (0.26, 0.95)	0.034	0.49 (0.24,0.99)	0.047
Palliative care – Yes	0.93 (0.45, 1.93)	0.84	-	-

^*Adjusting for age, sex, ethnicity, social work acuity, PICU LOS, severity

Respondents who were not Hispanic or Latino had a higher odds of high satisfaction (OR 2.09, 95% CI: 1.01, 4.33; p=0.047) compared to Hispanic or Latino respondents (Table 2). All 9 respondents who self-identified as Spanish-speaking only (n=2 with high satisfaction) were Hispanic/Latino; thus Spanish language was not used in the multivariate analysis. The Spanish-speaking respondents had lower odds of satisfaction than English speaking respondents (OR 0.14, 95% CI: 0.03, 0.71). Social work assessment of high acuity was significantly associated with satisfaction in both univariate analysis and in the multivariate analysis. In the multivariate analysis, patients identified as high acuity by social work had lower odds of high satisfaction (multivariate OR: 0.49, 95% CI 0.24, 0.99; p=0.047) (Figure 3).

Figure 3.

System and clinical categories of high and low satisfaction

Our initial hypothesized relationships with increased satisfaction, including older patient age, lower severity of illness and shorter LOS were not associated with increased satisfaction. Patient age was slightly lower in the higher satisfaction group at 6.7 years (SD 5.7) versus 7.3 years (SD 6.1) although this difference was not significantly associated with satisfaction (p=0.45). Similarly, higher severity of illness and shorter LOS tended to be higher in the high satisfaction group.

Event reports were filed on 32% (N=65) of patients; there was not statistical significance between patients who had event reports filed and their family’s satisfaction. In total, there were 128 free text comments (21% of free-text spaces for comments from all subjects). Over half of comments indicated areas for improvements or criticisms primarily focusing on communication with the care team, with the remainder indicating areas that went well.

Discussion

We identified a composite satisfaction level in our PICU and the subset of populations at greatest risk for low satisfaction. Our main findings are: 1) high-quality data with response rates of >90% collected in the PICU provides timely information on satisfaction, 2) one third of our population is not highly satisfied, and 3) we identified sub-populations at risk for low satisfaction, including those with Hispanic ethnicity and high social work acuity. These findings offer a framework to replicate for other PICUs and unit-level local improvement teams to learn of the needs of their population.

Our study piloted a process for obtaining high quality family satisfaction data in the PICU via a multi-faceted approach that incorporated family input from the start. Institutions looking to achieve similar high-quality data capture could allocate resources to follow a similar approach to garner this important information from patient families. The results of high capture and response rates suggest benefits of incorporating family preference during pre-implementation, using a PICU-specific survey, and delivering the survey in the PICU setting specifically. Response rates of greater than 90% greatly contrast with our standard process of obtaining Healthcare Consumer Assessment of Hospitals and Providers Surveys (HCAHPS) after discharge, which resulted in a total of only two completed surveys during 2017. Nationally, HCAHPS is the most widely used tool; while it offers measures of hospital quality by identifying common themes in feedback, response rates range from only 18-27% (13, 23). Beyond response rates, the approach in this project to use a population-specific survey adds to high data quality by asking specific questions specific to the context being evaluated. HCAHPS and other widely used surveys broadly measure experience of the entire hospital course (23). The process outlined for acquiring high-quality data is dependent on commitment to family satisfaction monitoring at the PICU leadership level.

It is not known what typical satisfaction rates are for families in the PICU. In studies of adult ICU patients, satisfaction rates range from 58 to 96% (14, 31). Our findings of overall unit satisfaction provide an opportunity not previously available to share these findings with frontline staff, physicians and advance practitioners, and unit leadership to target unit interventions to improve family and patient experience. The transparent reporting of meaningful metrics drives improvement, and awareness of unitwide satisfaction can provide a target for further improvement (32).

Intuitively, it is reasonable for PICU providers to link their view of successful care with an improving medical course for the patient. However, it has been demonstrated that patient and provider perspectives of successful care differ irrespective of medical outcome and are more dependent on open communication (12, 33). Appreciating comments from families, such as “the schedule felt more on the doctors’ schedule than ours” and “I felt rushed to ask questions during rounds before the doctor left”, may offer opportunity at the provider level to improve upon the current approach to family communication. Our results suggest a frank acknowledgement of what quantity of families are unsatisfied. Beyond recognizing a percentage of the population is unsatisfied, this process identifies exactly who they are and what system or patient variables are predictive of high or low satisfaction in our PICU. Understanding which variables are driving lower satisfaction offers PICU teams the opportunity to intervene with targeted improvement projects.

In our study, Hispanic families reported significantly lower satisfaction compared to non-Hispanic families. In contrast, in studies of adult outpatients, Hispanic populations have reported consistently high satisfaction (34). Access to interpreters may be one reason for lower satisfaction; a common theme among the free-text responses was frustration with the medical team only rounding once and the fear of missing rounds for communication with the team. This may be exacerbated in non-English speaking families. The availability of interpreters in stressful situations is comforting to families, even for families who report bilingual proficiency (35). In the last decade, there has been a rapid growth of Hispanic population in our state and referral region (36). While the increase in Hispanic population is not the same throughout the country, many referral hospitals are facing dynamically changing demographics that warrant attention when considering service provided to these patient populations (36). The ability to adapt to a continually changing patient population and their unique needs and perspectives necessitates on-going assessment to understand gaps and maintain high satisfaction.

Families with high social stressors correlated with lower satisfaction in our PICU. The impact of social stressors on inpatient satisfaction has not clearly been assessed. While somewhat intuitive that stressed families may have less satisfaction, this reflects a large population of patients in PICUs meriting further intervention to improve satisfaction. Patients of lower socioeconomic status (SES) are recognized to be at greater financial stress, have higher rates of post-ICU syndrome, and lack of follow-up after an ICU admission (37). Recent literature from Yagiela, and colleages, describe the resulting traumatic stress parents endure after pediatric critical illness, suggesting there is opportunity for improvement in the PICU setting to minimize long-term stress (38).

Limitations

There are limitations to this study. As a single-center study, our findings may not be generalizable to other PICUs. We did not study SES specifically, limiting the comparison for families with high social stressors. Additionally our population demographics with a fairly homogenous Caucasian population do not reflect many areas. Other hospitals’ infrastructure may inhibit the ease of approaching families in their room at time of discharge or transfer or not have the availability of the research associates. The social work acuity assessment, while standardized within our institution, is not validated; future studies using validated measures of social stressors would support further interventions to alleviate these stressors and ultimately improve satisfaction. Patients with LOS <48 hours were excluded to capture families fully exposed to PICU care. However, while families and patients with a short LOS will not have as many interactions with the care team, they represent nearly 40% of PICU patients locally and nationally, suggesting the importance of satisfaction studies in these patients (39). Finally, because of sensitivity to families whose child died in the PICU, we did not pursue administering surveys to this cohort; however future studies are warranted.

Conclusion

Future work to continue family satisfaction research is dependent on efficient data collection. This study offers a rigorous yet feasible approach to assess satisfaction with the clinical course in the PICU and begins to identify patient populations at risk for low satisfaction. In our population, Hispanic ethnicity and high social stressors at admission predicted low satisfaction, while illness severity, age and PICU LOS did not. Local improvement teams can use this approach to drive enhanced satisfaction. At the bedside level, there is substantial opportunity to feedback this data to providers for review of unit-level satisfaction and promote a proactive culture towards the needs of our populations. Finally, benchmarking of family satisfaction is beginning to take shape in the adult ICU population; a similar pathway could be followed in pediatrics (40). As the shape of healthcare continues to evolve and health institutions adapt to best match expectations of families, recognizing clinical and system drivers of inpatient satisfaction can provide opportunities to improve patient care.

Supplementary Material

supplementary 2

Supplemental Digital Content 2: Results from family Interviews assessing survey modality

supplementary 1

Supplemental Digital Content 1: Pediatric Family Satisfaction-ICU Survey

Abbreviations:

PICU

pediatric intensive care unit

LOS

length of stay

HCAHPS

Hospital Consumer Assessment of Healthcare Providers and Systems

pFS-ICU

parent family satisfaction in the intensive care unit

REDcap

Research Electronic Data Capture