Burden or Opportunity?

Abstract

Despite an ongoing need for pediatric research, low study enrollment may impede study completion, particularly in critical care. We conducted a prospective cross-sectional survey and chart review study to assess parent experiences with research in a pediatric intensive care unit (PICU). Of the 80 parents who completed the study survey, 54% were approached to participate in a research study in the PICU, and 93% agreed to participate. Motivators included altruism, low burden, low risk, and research that would benefit the child. Barriers included risks to the child, the child’s being too sick to participate, feeling overwhelmed, not having enough time to participate, the research’s being burdensome, and the research’s not being explained well. PICU parents had mostly favorable attitudes toward research participation. Compared with non-Hispanic survey respondents, respondents of Hispanic ethnicity less often reported having been approached for research, which further evidences the need to track research recruitment processes to help avoid underinclusion of members of minority populations.

Despite an ongoing need for pediatric research, low enrollment across studies may impede completion of research objectives.¹ Among parents who enroll their children and children who assent to enrollment, motivations for participation include research that provides benefit to the child, altruism, and familiarity with the care team and the research personnel.² Barriers include risks to the child, the added burden of the research in addition to clinical care or disruption to daily routine, and the timing and the way in which information about a study is presented to parents.³ Previous studies have found variable associations between socioeconomic factors and research participation in pediatric research.⁴

Enrollment may be particularly difficult in the critical care setting, where parents are likely to be experiencing stress because of their child’s condition. Fewer studies focus on factors relevant to the pediatric intensive care unit (PICU) specifically. In a prospective cohort study observing 271 consent encounters in pediatric critical care research, Menon et al. reported that consent rates were lower for cardiac surgery than other pediatric intensive care encounters and that consent rates were better when the research team was introduced by a member of the patient’s care team. Based on these results, they noted the need for more research into the influence of parental factors on consent rates.⁵ A subsequent qualitative study evaluated the perspectives of seven parents who agreed or declined to enroll their child in a research study while their child was a patient in one of two PICUs.⁶ The study found that factors related to willingness to enroll their child in research included (1) characteristics of the recruitment-consent encounter, such as timing, location, and information presented; (2) parental factors, such as emotional state, decision-making style, familiarity with the environment, past experiences, and personal motivation; (3) child characteristics, such as the child’s overall condition and response to interventions; and (4) study-related factors, such as risks, methods, burden, and benefit.⁷ The insights from this qualitative study demonstrated a need to explore the multiple factors relevant to participation in pediatric critical care research in a larger sample. The nature of research is also challenging in critical care settings, where a family may be approached by research teams of competing studies soon after their child’s admission to the critical care setting, when the patient’s clinical status may not be stabilized.⁸ Varying enrollment rates across multiple studies may result in strategies to increase enrollment such as approaching parents about multiple studies for which their child is eligible.⁹

We sought to build on prior work by investigating associations between reported factors related to research in critical care and noncritical care settings, with particular attention to the role of parental stress on research decisions and whether being approached for multiple studies affected parents’ willingness to enroll their child in a study. The objectives of our study were to (1) learn about the experiences of parents who are approached in the PICU to enroll their child in a research study, including what parents perceive to be research in the context of clinical care; (2) assess motivations and barriers for research participation; (3) evaluate associations between parental stress, child demographic and health characteristics, research characteristics, and concurrent eligibility for research and our primary outcome of research participation; and (4) evaluate associations between stress and barriers or motivations for research participation. We hypothesized that being approached multiple times about research studies, having a child with higher illness severity, and experiencing higher levels of stress would be associated with lower research participation.

STUDY METHODS

We conducted a prospective cross-sectional survey of parents and caregivers who were legal guardians (hereafter, parents and guardians are referred to as “parents”) of children admitted to the PICU and chart review study over a two-month period (July-August 2016) to assess experiences with research in the PICU in a large tertiary care academic medical center. This study period was chosen because the overall number of research studies recruiting participants in the PICU setting at that time was stable, thereby allowing us to assess parents’ perceptions about the simultaneous recruitment of participants for multiple studies. During this study period, all parents were also approached by PICU administrative staff members who gave them a modified consent form about enrolling their child in a PICU biobank. We sought to understand whether parents would characterize receiving the consent form as an attempt to recruit them to enroll their child in the biobank.

Survey development.

Parents completed a written survey that included questions about their experience with research, along four domains of interest (see figure 1). Survey questions were developed from analysis of the literature described above and discussion among the research team, which included individuals with expertise in qualitative research, health services research, statistics, and survey methodology. The survey was reviewed by a social worker and chaplain prior to administration. Questions were closed ended with predefined response options and were completed on paper. Questions asked about how individuals were approached for research, motivators and barriers to participate in research, and preferences for research participation. Parents also self-reported demographic information and whether they had previous experience with the PICU.

Figure 1.

Study Outcomes

Understanding parental stress due to a child’s acute illness is an evolving area without a standard for assessment. The survey assessed the parental experience of stress using the Severity of Acute Stress Symptoms—Adult scale,¹⁰ a seven-item Likert scale-based questionnaire assessing parental experience of stress-related symptoms during the preceding seven days that worsened following a stressful experience (e.g., PICU admission). Although this measure has not been previously validated in the population in our study, we selected it for its ability to measure acute stress due to recent traumatic experience, as we were interested in the experience of acute stress and parents’ willingness to enroll their child in research. Each of the seven items was scored on a scale of 0 to 4, where 0 reflected no agreement with the statement and 4 reflected complete agreement with the statement. The average total score was calculated according to published guidelines with a total possible score ranging from 0–28; a higher total score reflected a greater severity of stress.

Survey administration.

The study team screened for eligibility and identified parents of patients who were nearing readiness for discharge from the PICU. As this study did not require enrollment for clinical purposes during the acute phase of the patient’s illness, we chose to approach parents as the child was nearing clinical stability and thus discharge from the PICU to avoid additional stress during the acute phase of their child’s illness. Eligible parents included all parents whose primary language was English. Once participants were identified, the study team approached the parent(s) in the PICU at the bedside and provided them with an information sheet about the survey. Completing the survey indicated consent from those willing to participate. Once the completed surveys were returned, the study team immediately scored the stress measure, and any participant who scored greater than or equal to 3 on any question was referred to the PICU social worker for follow-up. In one instance in which a parent with a high score on the stress scale was referred to the social worker but had previously refused social work involvement, a referral was made to the PICU chaplain for follow-up. The surveys were deidentified and entered into the REDCap database by study personnel for analysis. A chart review was conducted to collect clinical information about the patient, including age, gender, length of PICU stay, the reason for admission, primary diagnosis, comorbid conditions, history of previous PICU admission, whether invasive interventions occurred, and insurance status. All study activities were approved by the hospital’s institutional review board. During the enrollment period, eight studies separate from this one were actively enrolling participants in the PICU for a range of observational and intervention studies.

ANALYSIS OF RESULTS

We conducted bivariate analyses to identify factors that influenced a parent’s willingness to enroll their child in research. As we knew that all parents were eligible for at least one study (the PICU biobank) and most participated in that study, we dichotomized participation between those who reported participation in zero to one study and those who reported participating in more than one study to evaluate enrollment in additional studies. We also assessed associations between stress and motivators or barriers for research participation and factors associated with being approached for research. Chi square and Fisher’s exact tests were used to measure associations between categorical variables. Continuous variables were checked for normality, and Wilcoxon rank sum was used to measure associations between ordinal or continuous variables and binary categorical variables. Kruskal-Wallis test was used to measure associations between nonbinary categorical variables and continuous variables. All tests were two-sided, and a P-value < 0.05 was considered statistically significant. Data were analyzed with SAS 9.4¹¹ and STATA 12.1.¹²

Eighty of 115 eligible parents were enrolled (response rate = 70%) and completed the study survey (see figure 2). The median age of parents was 37 years. The majority of participants were female, married, and primarily English speaking; had completed at least some college; and self-identified as white and non-Hispanic. More than one-half of the participants had previous experience with the PICU. The median length of stay for patients whose parents completed the study survey was two days (see table 1).

Figure 2. Study Enrollment¹.

¹ The participation rate was 70%.

Table 1.

Study Sample Characteristics

		Median	Interquartile range
Age range (years)		37	29, 44.5
PICU length of stay		2	1, 4.5
Acute Stress Scale score		4	0, 1
		N = 80	%
Sex	female	63	79
	male	17	21
Educational level	grade school high school or general	1	1
	equivalency diploma	12	15
	some college	26	33
	bachelor’s	25	31
	master’s or doctorate declined to provide or	14	18
	unknown	2	3
Primary language	English	74	93
	Spanish	2	3
	other	4	5
Ethnicity	Hispanic or Latino	18	23
	not Hispanic or Latino	59	74
	unknown	3	4
Race	white	56	70
	black	14	18
	Asian	2	3
	other	6	8
	unknown	2	3
Marital status	not married	25	31
	married declined to provide or	53	66
	unknown	2	3
Insurance type	public	27	34
	private	34	43
	both	14	18
	unknown	5	6
Comorbidity1	yes	29	36
	no	51	64
PICU intervention2	yes	62	78
	no	18	22
Previous PICU	yes	34	42
admission	no	46	58

^1.Comorbidities included developmental delay, premature birth, trisomy 21, cerebral palsy, CHARGE syndrome, attention deficit hyperactivity disorder, genetic disorders, and other medical disorders.

^2.PICU interventions included invasive mechanical ventilation, noninvasive mechanical ventilation, invasive line placement (venous or arterial), dialysis, vasoactive support, and cardiopulmonary resuscitation.

Parents’ experience with recruitment approach.

Of the 80 parents who completed the study survey, 45% indicated that they (or their child) were involved in research prior to this PICU admission. Despite the research team’s approaching all 80 parents with a modified consent form to enroll their child in a pediatric biobank, only 54% indicated that they were approached to enroll in any research during their child’s current admission to the PICU. However, of those who said they were approached at that time about any research (n = 43), 93% said they agreed to enroll their child in a study.

Of respondents who had been approached (during their child’s current PICU stay or a previous stay) about enrolling their child in research, 74% (56/76) reported that they felt respected while being approached, and about 67% (50/74) indicated that they had enough time to make a decision regarding participation. Most parents disagreed or strongly disagreed that they felt pressured (n = 56/73, 77%) or rushed (n = 54/76, 71%) to decide about participation or that they were approached too often (n = 54/76, 71%), about too many studies (n=53/76, 69%), or too close to admission (n = 51/76, 67%) or discharge (n = 47/75, 62%). When a child was eligible for multiple studies, one-third of parents preferred being approached by the treating doctor (n = 28/77, 36%), one-third preferred being approached by a single research staff person for all studies (n = 28/77, 36%), and one-fifth preferred being approached by a staff member from each study team (n = 17/77, 22%). Parents with research experience that predated the current PICU admission were more likely to prefer hearing about multiple studies from a single research staff person, while those without research experience were more likely to prefer hearing about multiple studies from the child’s treating doctor (Fisher’s exact p = 0.004).

Motivators and barriers regarding research participation.

Parents also reported motivators and barriers that influenced their decision regarding their child’s participation in research (figure 3). Of those who answered questions regarding motivations for participation, the most commonly reported motivators included that research would help future children (n = 32/38, 84%), was not burdensome (n = 26/32, 81%), was low risk (n = 25/31, 81%), and would help their child (n = 24/33, 72%). Barriers were overall less frequently reported, with the most common barriers being risks to the child (n = 10/17, 59%), feeling their child was too sick (n = 9/16, 56%), feeling overwhelmed by other things going on (n = 9/17, 53%), not having enough time (n = 8/16, 50%), the burden of the research (n = 8/17, 47%), and the research’s not being explained well enough (n = 6/15, 40%). Parents indicated that the research studies they were most interested in knowing about included studies that benefited their child (n = 53/63, 84%), benefited others (n = 47/57, 83%), or were recommended by their doctor (n = 41/56, 73%); and the majority said they would want to know about all research studies their child is eligible for (n = 39/56, 70%).

Figure 3.

Motivators and Barriers for Research Participation

Predictors of being approached for and participating in research.

We conducted bivariate analyses to identify factors that may have influenced parents’ willingness to enroll their child in research. All factors described in figure 1 were evaluated for associations with parental self-reporting of research participation. Participation in research was dichotomized to participation in zero to one study (n = 32) or participation in two to five studies (n = 11). Demographic factors, motivators and barriers, and factors related to the research approach were all considered as predictors of participation (see table 2). On bivariate analysis, only the number of times a parent was approached for research (Fisher’s exact p = 0.01) and the number of studies a parent was approached about (Fisher’s exact p < 0.0001) were significantly related to enrolling their child in research. Participation in prior research, knowledge of the PICU team, and the child’s illness severity as measured by the use of invasive PICU interventions were not significantly related to decisions to enroll in research. No motivators or barriers were significantly related to participation. On secondary analyses to evaluate associations between demographic factors and being approached about research participation, only ethnicity was significantly related to being approached, with individuals of Hispanic ethnicity less often reporting being approached to enroll their child in research (27% of Hispanic parents versus 62% of non-Hispanic parents, Fisher’s exact p = 0.02).

Table 2.

Factors Associated with Willingness to Participate in Research

Demographics and patient characteristics	P-value	Motivators and barriers	P-value	Research approach factors	P-value
Sex	1.00	Research helps my child	0.68	Number of times approached	0.01
Race	0.20	Research helps other children	1.00	Number of studies approached about	< 0.0001
Ethnicity	0.56	Research is low risk	1.00	Approached about too many studies	0.73
Language	1.00	Research is not burdensome	1.00	Approached about the right number of studies	0.23
Marital status	1.00	Research offers incentive	1.00	Approached about too few studies	0.96
Age	0.48	Doctor wanted me to do it	1.00	Approached too often	0.64
Educational level	0.78	Child wanted to do it	1.00	Approached the right number of times	0.25
Length of stay	0.59	Risk to child	1.00	Not approached enough	0.90
Acute stress score	0.20	Burden to child	1.00	Approached too soon after admission	0.25
Knowledge of PICU team	0.21	Not enough time	0.57	Approached at the right time	0.91
Previous PICU admission	0.31	Research not explained well	0.06	Approached too close to discharge	0.79
PICU interventions	1.00	Feeling overwhelmed	1.00	Felt respected in approach	0.25
Prior research participation	0.08	Child is too sick	0.53	Had ample opportunity to decide	0.35
Time commitment of prior research	0.28	Family does not want		Felt rushed to decide	0.72
		Could not do more	0.23	Felt pressured to decide	0.58
		Too many studies	0.40
		Lack of incentive	0.18

Parental stress and research participation.

We next assessed associations between parental report of stress, participation in research, and motivators or barriers for participation in research. Cronbach’s alpha for the Severity of Acute Stress Symptoms scale was 0.93. The sample median stress score was 4 (interquartile range [IQR] = 0, 11). Table 3 presents associations between stress scores and the reporting of motivators or barriers for research participation. Stress was not significantly related to parents’ decisions whether to enroll their child in research or to the reporting of any motivator or barrier to research participation.

Table 3.

Association between Stress Scores and Motivators and Barriers

Motivator	P-value	Barrier	P-value
Research helps my child	0.57	Risk to child	0.08
Research helps other children	0.11	Burden to child	0.31
Research is low risk	0.22	Not enough time	0.64
Research is not burdensome	0.53	Research not explained well	0.17
Research offers incentive	0.85	Feeling overwhelmed	0.47
Doctor wanted me to do it	0.35	Child too sick	0.60
Child wanted to do it	0.29	Could not do more	0.54
Any motivator	0.25	Too many studies	0.15
		Lack of incentive	0.11
		Any barrier	0.24

IMPLICATIONS FOR RESEARCH RECRUITMENT

Parents’ decisions regarding their child’s participation in pediatric critical care research may be influenced by a variety of factors related to the research process, the child, or the parent.¹³ We found, similarly to prior studies, that parents were motivated by the direct benefit to their child as well as altruism and were dissuaded by risks to their child, being overwhelmed, and their child’s health status.¹⁴ We sought to specifically explore the effect of being approached multiple times to enroll one’s child in research and parental stress as potential barriers to research participation.

We found a positive relationship between the number of times an individual was approached and the likelihood that the individual would enroll their child in research. This suggests that parents did not experience a burden that precluded participation by being approached multiple times for different research studies. However, as the overall rate of the parents’ children’s participation in research in the PICU setting was particularly high in this study, it is not clear if this finding would be generalizable to other settings. Alternatively, parents may have felt compelled to permit their child to participate if asked, although parents’ perspectives on feeling pressured about participation suggest this was not a primary driver. Additionally, although parents did not report feeling overburdened by being approached about multiple studies, the majority preferred being approached about all studies by a single point of contact (either a single research staff member or the treating doctor). This is not standard practice within many institutions, including ours, and this finding may inform recruitment practices in institutions with multiple studies occurring simultaneously.

While a prior study found parental anxiety as subjectively measured by an observer to be a barrier to research participation in the PICU,¹⁵ in our study, parent-reported stress as measured by the Acute Stress Disorder scale was not significantly related to parents’ decision about enrolling their child in research. Several reasons may account for this finding. First, the measure used to assess acute stress has not previously been validated in the PICU population enrolled in our survey study. Additionally, the range of scores for parents in our study sample was relatively narrow, and the median stress score was at the low end of the range, even though parents reported that being overwhelmed may have negatively affected their decision whether to enroll their child in research. The measure, which was designed to assess the clinical symptoms of acute stress, may have failed to pick up on other symptoms of stress, including the feeling of being overwhelmed. Moreover, we measured parent stress as children were preparing to leave the PICU, which was likely a lower period of stress than the time of PICU admission. This may not have affected responses for parents of children with shorter lengths of stay, as the measure assesses symptoms over the prior week; however, for parents of children with longer lengths of stay, this might have led to lower reported stress. Additionally, as noted in the limitations we describe below, our sample size may not have been adequate to identify a relationship between parental stress and research participation or barriers to research participation.

We found that almost one-half of the study sample reported not being approached about research in which their child might be eligible to participate. This finding was significant, as all parents who were screened for inclusion in our study had been approached by an administrative person at the time of PICU admission about the PICU biobank. This finding suggests that not all parents recognized the PICU biobank as research or not all recalled receiving or completing the biobank consent form. Alternatively, parents may not have recognized distribution of a research consent form by administrative staff members as communicating a research opportunity, compared with an approach by a staff person whose sole role is to recruit research participants. The finding may have implications about how parents perceive what constitutes research and the potential effect of this perception on their decisions about their child’s participation. Additionally, we found that parents of Hispanic ethnicity were less likely to report being approached about research participation, which is consistent with other reports of racial and ethnic differences in rates of being approached for research in the PICU.¹⁶ This finding thus further evidences the need to track research recruitment processes to avoid underinclusion of minority populations.

Several important limitations about our study must be noted. First, this was a single-center study, reflecting parental experience in the context of research being done over a discrete period. As context may have an influence on parents’ experience about research, our findings might not apply to a setting with a different process for recruiting and enrolling research participants or a different mix of interventional and observational studies. Second, there was a lack of variation in our primary outcome of research participation due to the high reported rate of participation, which made it challenging to identify any associations between potential predictors and research participation. For example, we did not find associations between research participation and socioeconomic status markers (race or ethnicity, insurance status, and educational level) or health status, which have been previously reported.¹⁷ This could be explained by the inadequate power to detect this association or could reflect the homogeneity of the study population, as we did find that Hispanic individuals less often reported being approached about enrolling their child in research and therefore could not contribute information or perceptions about their experience with research in this survey. Third, as we collected information on research participation based on parents’ self-reporting, recall bias may have influenced responses. We hoped to minimize this by asking primarily for information regarding the child’s current PICU admission. Fourth, although our survey was designed with expert input, it was not formally validated, and the design limited the opportunity for follow-up questions to clarify areas of uncertainty. Fifth, the Severity of Acute Stress Symptoms—Adult was not validated in our study population. Finally, our study excluded non-English-speaking families because the survey instrument was written in English, which may also limit the generalizability of our findings.