Attitudes Towards COVID-19 Vaccine Among Pediatric Patients with Sickle Cell Disease and their Caregivers

Abstract

Objective:

To evaluate attitudes towards vaccination and vaccine uptake regarding coronavirus disease 2019 (COVID-19) among pediatric patients with sickle cell disease (SCD) and their caregivers.

Procedure:

Adolescent patients and caregivers of children with SCD, were surveyed during routine clinic visits; we then conducted a logistic regression analysis to understand differences in vaccine status while qualitative responses were coded thematically.

Results:

Among respondents, the overall vaccination rate among adolescents and caregivers was 49% and 52%, respectively. Among the unvaccinated, 60% and 68% of adolescents and caregivers, respectively, preferred to remain unvaccinated, most commonly due to lack of perceived personal benefit from vaccination or mistrust in the vaccine. Multivariate logistic regression analysis showed that child’s age (Odds Ratio [OR]=1.1, 95% Confidence Interval [CI] 1.0–1.2, P<0.01) and caregiver education (measured by the economic hardship index (EHI) score, OR=0.76, 95% CI, 0.74–0.78, P<0.05) were independent predictors of getting vaccinated.

Conclusion:

Despite the increased risk of severe illness due to COVID-19 in patients with SCD, vaccine hesitancy remains high in this population of families whose children have SCD. Fortunately, the reasons cited for deferring vaccination among those who are unvaccinated were largely due to barriers that may be overcome with quality communication around the utility of the vaccine and information about vaccine safety.

INTRODUCTION

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection and the resulting coronavirus disease 2019 (COVID-19) have resulted in serious morbidity and mortality since its emergence in 2020. Immunizations are the most effective intervention for reducing severe disease and mortality, particularly among individuals with certain chronic medical conditions such as sickle cell disease (SCD). Adults and children with SCD with COVID-19 are at increased risk of developing severe vaso-occlusive (pain) crises, acute chest syndrome, nephropathy, abnormal transcranial dopplers, and strokes.¹.

In the United States, historically underrepresented minorities (URM) have poor vaccine uptake, especially among older adolescents and adults.² Over 90% of patients with SCD identify as URM.³ However, the overall vaccination rates for COVID-19 among Black and Hispanic individuals in the United States have lagged behind Asians and Non-Hispanic Caucasians.⁴ With this in mind, we sought to assess actual vaccine uptake for COVID-19 among adolescents with SCD and household caregivers of children with SCD, all of whom identified as URM. To identify strategies to address vaccine hesitancy, we sought to broadly identify attitudes and barriers among unvaccinated adolescents and pediatric SCD caregivers.

METHODS

From May 2021 to February 2022, we surveyed a convenience sample of caregivers of patients with SCD (any genotype) aged 5–12.9 years and adolescents aged 13–18 as part of a larger institutional study of caregivers of all patients with SCD and affected adolescents ≥ 13. We first asked whether the vaccine-eligible adolescent or caregiver had received at least one dose of the COVID-19 vaccine. If not, we then inquired about plans and attitudes towards vaccination. Using the consensus coding strategy,⁵ we pooled reasons for declining vaccines into three main categories as shown in Fig. 1C. Using information from the St. Jude Sickle Cell Clinical Research and Intervention Program (SCCRIP) Database participant demographic information (age, sex, race, ethnicity, genotype, economic hardship index (EHI)⁶, parental education level, household income) were examined for potential associations with vaccination status. The St. Jude IRB approved this study.

Figure 1 –

Results from survey of adolescent patients with Sickle Cell Disease (SCD) and caregivers of patients with SCD

Statistical analysis:

Two-sample t-test or Wilcoxon rank sum test and Fisher’s exact test were used to evaluate two group differences on demographic variables. The Shapiro–Wilk test was used to test for normality of the data. Univariate logistic regression model was used to assess the associations of demographic variables with vaccination status by caregiver and adolescents, separately. Covariates with significant associations at P < 0.10 were included in the multivariable logistic regression analyses to obtain the final model using stepwise model selection. All the covariates were tested for multicollinearity before entering the model (a variance inflation factor [VIF] <2). For adolescents, we have three pairs of siblings whose vaccination status are same. So, we did a sensitivity analysis by removing one sibling from each pair for adolescent analysis to avoid correlations and the conclusions are similar and omitted. All P-values were two-sided. All analyses were performed in R version 4.1.1.

RESULTS

Demographics and household status of respondents (132 caregivers, 49 adolescents) are shown in Table 1. Approximately half of the caregivers (48%) and adolescents (51%) were unvaccinated (Figure 1a) and most of this group expressed no intention to get vaccinated (Figure 1b). Respondents, who were unvaccinated and did not intend to get vaccinated, cited a lack of a perceived benefit to vaccination (caregivers 45%, adolescents 53%), a lack of confidence or trust in the vaccine (caregivers 29%, adolescents 40%), and less commonly, unanswered medical questions (caregivers 12%, adolescents 7%) as reasons for choosing not to get vaccinated (Figure 1c) along with exemplar quotes shown in Figure 1d. In comparisons across all demographic variables, more educational hardship as measured by EHI (P=0.04) and younger child age (P=0.01) were observed in caregivers who did not receive vaccine compared to those who received vaccine, whereas for adolescents increases in hardship (as measured by all tested EHI variables) were observed in adolescent who did not receive compared to those who received vaccine (P ≤ 0.02) (Table 1).

TABLE 1.

Participant demographic data and household demographic status

	Caregivers (N=132)			Adolescent patientsa (N=49)
	N=132	No (n=63)	Yes (n=69)	p-valueb	No (n=25)	Yes (n=24)	p-valueb
Participant Demographic by Vaccine Status
Gender, n (%)				1.00			1.00
Female	69 (52)	33 (52)	36 (52)		13 (52)	12 (50)
Male	63 (48)	30 (48)	33 (48)		12 (48)	12 (50)
SCD Type, n (%)				0.25			0.26
SB+	9 (7)	3 (5)	6 (9)		2 (8)	1 (4)
SC	34 (26)	13 (21)	21 (30)		4 (16)	9 (38)
SS/SB0	89 (67)	47 (74)	42 (61)		19 (76)	14 (58)
Patient Race, n (%)				1.00			0.49
Black	131 (99)	63 (100)	68 (99)		25 (100)	23 (96)
Other	1 (1)	0(0)	1 (1)		0 (0)	1 (4)
Child’s Age				0.01			0.60
Mean (SD)	9.5 (5.1)	8.4 (5.1)	11 (4.9)		15 (2.7)	16 (1.4)
Median (Range)	9.4 (1.2~19)	7.9 (1.2~18)	11 (1.4~19)		15 (5.1~19)	16 (13~18)
Enrolled on SCCRIP, n (%)				0.54			0.61
No	11 (8)	4 (6)	7 (10)		1 (4)	2 (8)
Yes	121 (92)	59 (94)	62 (90)		24 (96)	22 (92)
Household Demographics by Vaccine Status
Average Household Size				0.22			0.06
Mean (SD)	6 (3)	5.4 (2.9)	6.4 (3.1)		5.8 (3)	7.9 (2.8)
Median (Range)	5 (2~10)	4 (2~10)	5.5 (2~10)		4 (3~10)	10 (3~10)
Caregiver's Education Level, n (%)				0.24			0.07
College Graduate	19 (41)	6 (27)	13 (54)		2 (22)	4 (80)
GEDc	4 (9)	1 (5)	3 (13)		1 (12)	0 (0)
Less than a high school degree	6 (13)	4 (18)	2 (8)		0 (0)	1 (20)
Received a high school diploma	11 (24)	7 (32)	4 (17)		3 (33)	0 (0)
Some college	6 (13)	4 (18)	2 (8)		3 (33)	0 (0)
EHI Overalld				0.07			0.01
Mean (SD)	28 (14)	31 (14)	26 (13)		28 (12)	21 (14)
Median (Range)	28 (10~62)	31 (10~62)	26 (10~52)		28 (10~50)	17 (10~52)
EHI Povertyd				0.12			<0.01
Mean (SD)	23 (18)	26 (19)	21 (17)		27 (17)	15 (14)
Median (Range)	18 (1~71)	22 (2~71)	10 (1~61)		25 (4~69)	10 (3~48)
EHI Educationd				0.04			0.02
Mean (SD)	23 (16)	26 (18)	20 (13)		27 (15)	17 (14)
Median (Range)	20 (3~100)	24 (3~100)	14 (3~55)		25 (5~61)	10 (3~51)

Abbreviations: SB+, Sickle Beta-Plus Thalassemia; SC, Sickle Hemoglobin- C Disease; SS, Sickle Cell Anemia; SBO, Sickle Hemoglobin-O Disease; SD, Standard Deviation.

^aAmong adolescent patients, there were three pairs of siblings, and each of them had the same vaccine status.

^bp-values <0.05 are in bold. Fisher’s exact test was used to compare gender, SCD Type, Patient Race, Enrolled on SCCRIP, and Caregiver's Education Level. Wilcoxon rank-sum test was used to compare the Child’s age, Average Household Size, EHI Overall, EHI Poverty, and EHI Education between the Caregivers and Adolescents groups.

^cGED denotes a General Educational Development Certificate, which is awarded in lieu of a high school diploma.

^dEHI denotes economic hardship index: a validated measure of neighborhood socioeconomic status standardized on a scale from 0–100 in which higher scores indicate higher hardship.⁸ The education subscale is based on the percentage of individuals in a community with no high school diploma or equivalent, while the poverty subscale is the percentage below the poverty level.

In univariate logistic regression analyses to assess the associations of demographics and household status with vaccine status, we found child’s age, overall EHI, caregiver’s education level, and EHI education to be associated with caregiver vaccine status at P< 0.1 (Table 2). EHI overall and caregiver’s education level were excluded from the multivariate model analysis because of their strong correlation with EHI education (VIF>2). Multivariate logistic regression analysis showed that child’s age (Odds Ratio [OR]=1.1, 95% Confidence Interval [CI], 1.0–1.2, P<0.01) and EHI education (for 10-unit increase: OR=0.76, 95% CI, 0.74–0.78, P<0.05) are the independent predictors of getting vaccinated in caregivers. For adolescents, univariate regression analyses showed that average household size, EHI overall, EHI poverty, and EHI education are associated with vaccine status at P < 0.1 (Table 2). Due to correlations among EHI variables and average household size (VIF>2), EHI for poverty is the only significant predictor of getting vaccinated (for a 10-unit increase in EHI: OR=0.57, 95% CI, 0.54–0.60, P=0.02). After removing one sibling from each pair of siblings, sensitivity analysis showed the same conclusions.

TABLE 2.

Univariate logistic regression analysis for assessing associations of vaccine status by caregiver and adolescents

	Caregivers		Teens
	OR (95% CI)a	p-valueb	OR (95% CI)a	p-valueb
Participant Demographic
Gender		0.98		0.89
Female	Reference
Male	1 (0.5–2)		1.1 (0.4–3.3)
SCD Type		0.23		0.22
SB+	Reference
SC	0.8 (0.2–3.8)	0.79	4.5 (0.3–65)	0.27
SS/SB0	0.5 (0.1–1.9)	0.28	1.5 (0.1–18)	0.76
Patient Race		0.99		0.99
Black	Reference
Other	> 99 (0-Inf)		>99 (0-Inf)
Child’s Age	1.1 (1.0–1.2)	0.005	1.0 (0.8–1.4)	0.80
Enrolled on SCCRIP		0.43		0.54
No	Reference
Yes	0.6 (0.2–2.2)		0.5 (0.1–5.4)
Priori family information from the SCRIPP database
Average Household Size	1.1 (1.0–1.3)	0.18	1.3 (0.1–1.7)	0.07
Caregiver's Education Level
College Graduate	Reference
GEDc	1.38 (0.12–16.2)	0.80	0 (0-Inf)	1
Less than a high school degree	0.23 (0.03–1.63)	0.14	>99 (0-Inf)	1
Received a high school diploma	0.26 (0.06–1.26)	0.01	0 (0-Inf)	1
Some college	0.23 (0.03–1.63)	0.14	0 (0-Inf)	1
EHI Overalld	0.75 (0.73–0.77)	0.06	0.59(0.56–0.63)	0.045
EHI Povertyd	0.85 (0.83–0.87)	0.17	0.57(0.54–0.60)	0.02
EHI Educationd	0.76 (0.74–0.78)	0.047	0.58(0.56–0.61)	0.03

Abbreviations: OR, Odds Ratio; CI, Confidence Interval; Inf, Infinity.

^aOdds Ratios and 95% Confidence Intervals were calculated based on a 10-unit increase for three EHI variables. Reference is the reference category of the variable.

^bp-Values<0.05 are in bold.

^cGED denotes a General Educational Development Certificate, which is awarded in lieu of a high school diploma.

^dEHI denotes economic hardship index: a validated measure of neighborhood socioeconomic status standardized on a scale from 0–100 in which higher scores indicate higher hardship.⁸ The education subscale is based on the percentage of individuals in a community with no high school diploma or equivalent, while the poverty subscale is the percentage below the poverty level.

DISCUSSION

Despite the increased risk of serious morbidity associated with COVID-19 among individuals with SCD, and the ability of the vaccine to decrease some of this risk, COVID-19 vaccine uptake was suboptimal among adolescent patients and household caregivers of children with SCD. This low uptake of the COVID-19 vaccine appears to be consistent with suboptimal uptake of the seasonal influenza vaccine among patients with SCD.^2,7 At the time of survey collection, younger children were not vaccine-eligible outside of a clinical trial, hence younger children were not included in this study. Nevertheless, we have not observed changes in vaccine attitudes among clinic patients since the approval of COVID-19 vaccines for younger children This low vaccine uptake for COVID-19 occurred even though vaccination was offered as part of our standard of care in SCD clinic, and contrasts with the acceptance of other recommended vaccines such as the PCV and MCV vaccines where vaccine rates in our clinic are 75% and 76% respectively.

The positive association with lower education attained, as measured by the EHI subscale, is consistent with published data demonstrating higher vaccine hesitancy among those who did not complete high school.⁸ Exploration of a caregiver’s educational attainment and knowledge of economic hardship may help identify families at increased risk of declining vaccination and prepare providers to allow time for additional counseling.

Limitations of this study include responses drawn from a single urban center in the central-southeastern United States and as survey responses were analyzed asynchronously by the nursing research team outside of clinic visits, we were unable to collect data on post-vaccine counseling by hematology clinic staff. While we are relatively confident in our vaccination rates for the PCV and MCV vaccines, which are primarily offered in our clinic to patients with SCD, we have incomplete data on uptake of the seasonal-influenza vaccine. We did not query families as part of this survey and as many families choose to receive the influenza vaccine in the community, the data recorded in our electronic medical record is likely incomplete and not reflective of true vaccination rates. Finally, although children aged 12 were vaccine-eligible at the time of survey collection, the parent survey enrolled only children ≥ 13 years, and thus 12-year-olds were excluded from this study.

In clinical practice, vaccine hesitancy by patients or caregivers may be overcome by simply framing vaccination as the default clinical option, versus offering patients a choice.⁹ In other words, directly offering vaccination as part of the care plan requiring the patient (or caregiver) to actively opt-out may be a better option, instead of asking, “Would you like your child to receive the COVID-19 vaccine today?” which requires the decision-maker to opt-in to vaccination. Clinicians may find the American Academy of Pediatrics learning collaborative models for framing vaccines as usual care to be helpful.^10,11 This presentation of vaccination as the default option frames the choice in such a way that it is likely to trigger unconscious reasoning towards acceptance of the particular choice – borrowed from behavioral economics, this is known as a ‘nudge’. Simple measures, such as the nudge, may help boost COVID-19 vaccination rates in patient populations.^12–14

If a patient or caregiver expresses hesitancy after this presumptive nudge, we recommend that providers explore the expressed hesitation in a non-judgmental manner to understand the reasons underlying the refusal. For example, while most unvaccinated respondents indicated little interest in future vaccination, probing the reason behind their decision yielded insightful and actionable information. The most common barrier identified among hesitant unvaccinated individuals was a lack of perceived benefit to vaccination, followed by lack of confidence in the vaccine, and unanswered medical questions. These are all potentially modifiable barriers to vaccine hesitancy, that open and effective communication may overcome. For example, the participant stating, “I am scared I’ll catch COVID if I get the shot,” might reconsider if a trusted provider explained why the vaccine is unable to cause disease. Recognizing that educational attainment and economic hardship are predictive of refusal, it is particularly important to use lay terminology and avoiding medical jargon. Patients or caregivers may have limited health literacy and benefit from a simple framing about the history of vaccine development that informed the COVID-19 vaccine (especially for those who believe the vaccine science was rushed). This communication about vaccine utility extends beyond COVID-19 and may be useful for other recommended vaccines such as the seasonal influenza vaccine.

For those who remain vaccine averse despite provider nudging, curiosity, and education, other strategies such as ongoing bi-directional communication about vaccination, input from trusted community members or other patients, incentives or other forms of “nudges” may result in reconsidering a previous refusal.¹⁵ In families who continue to decline a recommended vaccine or booster, it is important that providers respect the decision and continue to identify determinants of vaccine hesitancy. Providers should periodically revisit the topic in a supportive and empathetic manner: “nudges” can be an ongoing process.

Abbreviations:

SARS-CoV-2

severe acute respiratory syndrome coronavirus 2

COVID-19

coronavirus disease 2019

SCD

sickle cell disease

URM

underrepresented minorities

EHI

economic hardship index

VIF

variance inflation factor