Influenza-Related Knowledge, Beliefs, and Experiences Among Caregivers of Hospitalized Children

In this cross-sectional survey study, English- and Spanish-speaking caregivers were eligible for participation if their child was aged 6 months through 17 years and hospitalized on a medical, surgical, rehabilitation, or psychiatric unit at a tertiary care pediatric hospital between November 2017 and April 2018. Caregivers were excluded if they completed the survey previously. Respondents were drawn from participants in the hospital’s survey-based outcomes assessment program (OAP), which collects parent- and patient-reported outcomes data from eligible, consenting families of hospitalized children (Supplemental Figure 2). This study was approved by the hospital’s institutional review board. An exempt determination was made by the state’s institutional review board regarding review of vaccine registry data.

Since 2008, the hospital has conducted automated EHR-based screening to (1) identify influenza vaccine-eligible patients based upon caregiver report, hospital system records, and/or verification in the state vaccine registry; (2) prompt providers to offer influenza vaccine; and (3) document caregiver acceptance or declination (including declination reason).⁸  Community influenza activity was more severe in the 2017–2018 season than previous seasons, with peak activity in late January to early February.¹³ 

The survey instrument was developed by using an iterative, systematic approach. First, we conducted a targeted review of published surveys assessing factors relevant to pediatric influenza vaccination and evaluated these using established criteria.¹⁴  Forty-four items in 10 key content areas were identified, adapted for the hospital setting, and reviewed by 6 national vaccine experts. On the basis of expert agreement regarding items that most fully captured each content area, 20 final items addressing influenza-related knowledge, beliefs, and experiences were selected (Supplemental Table 4). Most response options were based on a 4-point Likert scale (ie, strongly agree, agree, disagree, and strongly disagree). The instrument also included 12 items from the Parent Attitudes about Childhood Vaccines (PACV) survey, modified and validated for use with influenza vaccine.^15,16  These items were associated with previous influenza vaccine declination during hospitalization.¹⁰  The survey also addressed useful information for vaccine decision-making during hospitalization and preferred information sources. Final English and Spanish versions were embedded within the hospital’s OAP admission survey.

Between November 2017 and April 2018, the OAP team identified, approached, and obtained consent from families eligible for the OAP within 72 hours of hospital admission. They used EHR data to identify OAP participants also eligible for the influenza survey. These families completed the survey in Research Electronic Data Capture via self-administration or via telephone with the OAP team. Survey responses were later linked to demographic, clinical, and visit information abstracted from the child’s EHR. For vaccine-eligible children whose caregivers declined influenza vaccine during hospitalization and were not seen subsequently in the hospital system, vaccine registry data were obtained to capture doses administered elsewhere posthospitalization. An estimated ≥93% of children and adolescents in the state participated in the registry during the study period.¹⁷ 

The main outcome measures were child influenza vaccination status at hospital admission, assessed among all respondents, and caregiver acceptance of influenza vaccine for their child during hospitalization, assessed in a subgroup of respondents with influenza vaccine-eligible children (ie, no vaccine contraindications and not yet vaccinated this season). Child influenza vaccination status at hospital admission, dichotomized into already vaccinated versus not yet vaccinated this season (ie, between August 1, 2017, and hospital admission), was determined by using EHR-documented responses to the hospital-wide influenza vaccination screening program or, if unavailable, EHR or state vaccine registry data. Caregiver acceptance of influenza vaccine for their child during hospitalization was determined by using screening responses.

Caregiver-reported knowledge, attitudes, experiences, and vaccine hesitancy served as the main independent variables. We dichotomized response options using a 4-point Likert scale (ie, agree versus disagree). Vaccine hesitancy was determined by collapsing each PACV item response into 3 categories (nonhesitant, not sure, and hesitant) and assigning a score of 0 (nonhesitant), 1 (not sure), or 2 (hesitant), converting the total raw score into a 0- to 100-point scale by using simple linear transformation that accounted for missing data, and dichotomizing this score into <50 (nonhesitant) and ≥50 (hesitant).^10,15,16  Other independent variables included sociodemographic characteristics, child chronic disease status, hospital service, and hospitalization duration and timing. Chronic disease status was categorized as complex chronic disease, noncomplex chronic disease, or no chronic disease by using the validated Pediatric Medical Complexity Algorithm.¹⁸  Hospital service was categorized as medicine, surgery, psychiatry, or rehabilitation. To address time-dependent variability in influenza vaccination, hospitalization timing was categorized as late fall (November 29, 2017, to January 14, 2018), winter (January 15 to February 28, 2018), and spring (March 1 to April 29, 2018).

Sociodemographic, clinical, and visit characteristics; influenza vaccination status at admission; influenza vaccine acceptance during hospitalization; influenza vaccination posthospitalization; and survey responses were summarized by using descriptive statistics. Multivariable logistic regression models were used to examine individual associations between (1) caregiver survey responses and their child’s influenza vaccination status at admission (all respondents) and (2) caregiver survey responses and influenza vaccine acceptance for their child during hospitalization (subgroup of respondents with vaccine-eligible children). Models were adjusted for respondent age, race and/or ethnicity, education, relation to child, child age, insurance, chronic disease status, and hospitalization timing; hospital service and hospitalization duration were also included in models examining vaccine acceptance during hospitalization. Covariates were selected a priori on the basis of their potential association with influenza vaccine decision-making and after assessing for multicollinearity. Language was not included in the models because of instability, likely reflecting the small sample of non-English speakers (n = 12). Standardization of variables in the regression models was performed. To account for multiple comparisons, P values were adjusted for false discovery rate on the basis of the Benjamini-Hochberg method.¹⁹  We performed 2 sensitivity analyses. First, we excluded patients admitted to the rehabilitation or psychiatry unit (n = 7). Second, in the multivariable models assessing vaccine acceptance during hospitalization, we limited covariates to variables associated with the outcome at P < .10 in bivariate analysis (parent race and/or ethnicity, education, child age, and hospitalization timing). Our findings remained unchanged; thus, we retained all patients and covariates in the final models. Analyses were conducted by using SAS version 9.4 (SAS Institute, Inc, Cary, NC).

Of 587 eligible families, 522 (88.9%) completed the survey, representing 8.9% of the hospital population during the study period (Supplemental Fig 2). Most respondents were non-Hispanic white, English-speaking, and/or college-educated (Table 1) (Supplemental Table 5). Most children were aged 5 through 17 years, privately insured, and/or hospitalized on the medicine service; nearly three-quarters had underlying chronic disease. Demographic characteristics did not differ between eligible caregivers who did versus did not complete the survey; a lower proportion of the hospital population were English-speaking (90.2%), non-Hispanic white (51.4%), and privately insured (53.0%). At hospital admission, 63.2% (95% confidence interval [CI]: 59.0%–67.2%) of study patients had already been vaccinated in the 2017–2018 season (Fig 1). By the season’s end, 79.4% (95% CI: 75.6%–82.7%) were vaccinated.

Most respondents agreed that children are at risk for influenza infection (95.5%) and can experience adverse complications (93.9%) (Table 2). Approximately three-quarters felt that influenza vaccination is important for everyone’s health and their child’s health specifically. Most agreed that the vaccine is effective (67.9%) and safe (86.5%); fewer (18.3%) worried that their child could get the flu from the vaccine. One in 5 respondents were identified as vaccine hesitant (median PACV score: 20.8; interquartile range: 8.3–41.7). Most respondents knew that influenza vaccine is recommended for their child (84.5%) and reported receiving a strong recommendation from their child’s primary care provider (90.6%). Approximately half agreed it is important for their child to receive influenza vaccine at their primary care provider’s office. Two-thirds believed a child with mild illness should wait to receive the vaccine.

In bivariate analysis, caregiver knowledge about influenza vaccination, beliefs regarding influenza infection and vaccination, receipt of a strong primary care provider recommendation, and previous influenza vaccination were positively associated with having an already vaccinated child at admission (Table 2). Caregiver vaccine hesitancy, lack of knowledge about vaccine side effects and vaccination timing, and a child’s previous adverse vaccine reaction were negatively associated. In multivariable analysis, the caregiver belief that influenza vaccination is important for their child’s health was the strongest positive predictor of having an already vaccinated child at admission, whereas vaccine hesitancy was the strongest negative predictor.

Overall, 178 respondents (34.1%) had a child who was influenza vaccine-eligible at hospital admission (Fig 1). Of these, 30.3% (95% CI: 24.1%–37.5%) accepted and 65.7% (95% CI: 58.5% –72.3%) declined influenza vaccine for their child during hospitalization; 3.9% (95% CI: 1.9%–7.9%) had no vaccine decision documented in their child’s EHR (Table 3). EHR-documented declination reasons are presented in Supplemental Table 6. Of children whose caregivers declined influenza vaccine during hospitalization and had known vaccination status posthospitalization (n = 110), 15 (13.6%) had documented receipt by the season’s end. Of those who declined at hospital screening because they wished to wait, discuss further, or receive it elsewhere (n = 33), 10 (30.3%) had documented receipt by the season’s end.

In bivariate analysis, caregiver knowledge about influenza vaccination, beliefs regarding influenza infection and vaccination, receipt of a strong primary care provider vaccine recommendation, and previous influenza vaccination of their child were positively associated with influenza vaccine acceptance during hospitalization (Table 3). Caregiver vaccine hesitancy and agreement that influenza vaccination should be deferred during acute illness were negatively associated with vaccine acceptance. In multivariable analysis, caregiver views regarding the importance of influenza vaccination were the strongest positive predictors of vaccine acceptance during hospitalization. Caregiver vaccine hesitancy and agreement that vaccination should be deferred during acute illness were the strongest negative predictors. In adjusted analyses, caregiver-reported importance of getting the vaccine with their child’s primary care provider was also negatively associated with influenza vaccine acceptance.

Among respondents with vaccine-eligible children, 87 (48.9%) indicated that the following information would help influenza vaccine decision-making during hospitalization (>1 selection permitted): influenza vaccine (n = 44; 50.6%), influenza infection (n = 37; 42.5%), when their child should receive the vaccine (n = 29; 33.3%), and why their child should be vaccinated (n = 28; 32.2%). Preferred information sources included discussion with a doctor (n = 37; 42.5%) or nurse (n = 30; 34.5%), paper handouts (n = 33; 37.9%), and online information (n = 20; 23.0%).

In this study, we used a rigorously designed survey instrument to examine caregiver knowledge, attitudes, and experiences associated with influenza vaccine decision-making for hospitalized children. The findings offer insight into a range of influencing factors not described previously in this setting. Specifically, caregiver views regarding the health benefits of influenza vaccination, knowledge about vaccination timing, and vaccine hesitancy strongly influenced vaccine decision-making before and during hospitalization. Two-thirds of study families with influenza vaccine–eligible children declined influenza vaccine during hospitalization; only 14% of these children were vaccinated by the season’s end. This finding is worrisome because many hospitalized children have underlying conditions that increase their risk for influenza infection and related complications,^2,20–22  and vaccination opportunities are often missed in other settings.^23,24  Evidenced-based interventions are needed to increase influenza vaccination in this vulnerable population. This study, by elucidating key factors impacting caregiver vaccine decision-making and characterizing their informational needs and preferences, may inform this future work.

Two-thirds of study children were already vaccinated at hospital admission, a higher proportion than the 37% to 51% reported previously in pediatric inpatient populations.^8,9,11  Conversely, only one-third of caregivers with influenza vaccine–eligible children accepted influenza vaccine during hospitalization, in contrast with studies showing that approximately half of families accept the vaccine.^8–10  These discrepancies may reflect our late-fall initiation of recruitment, with many families having already accepted the vaccine at previous health care encounters (“early adopters”) and a higher proportion of remaining families resisting influenza vaccination. Supporting this, influenza vaccine acceptance was lowest among caregivers with children admitted in the spring (Supplemental Table 5). For these high-risk families, targeted educational interventions may be beneficial.²⁵ 

In this study, caregiver views regarding the importance of influenza vaccination were key predictors of influenza vaccination before and during hospitalization. The belief that influenza vaccination offers the best protection against disease also had a strong positive effect on vaccine decision-making. These findings emphasize the importance of promoting the health benefits of influenza vaccination for hospitalized children. Other factors, such as caregiver knowledge regarding influenza vaccine safety, effectiveness, and recommendations; their child’s perceived infection risk; and previous caregiver vaccination also positively impacted vaccine decision-making. An earlier survey similarly revealed an association between parental agreement that influenza vaccine is safe and recommended annually and child vaccination status at hospital admission.¹¹  Our findings also align with survey results from nonhospital settings.^26–33 

In the current study, we identified 20% of caregivers as vaccine hesitant and found a strong negative association between vaccine hesitancy and influenza vaccine decision-making before and during hospitalization. These data are consistent with findings from a retrospective study of parents with previously hospitalized children¹⁰  and a study conducted prospectively in the pediatric emergency department.¹⁶  They highlight the need to address vaccine hesitancy in high-risk populations receiving care in diverse health care settings.

Three-quarters of respondents agreed that influenza vaccination should be deferred during mild acute illness, and these caregivers had markedly lower odds of vaccine acceptance during hospitalization. This concern about concurrent illness, consistent with a previous survey,¹²  could reflect a desire to avoid additional discomfort for their child during acute illness. It also may indicate a lack of understanding regarding true influenza vaccine contraindications. According to Advisory Committee on Immunization Practices guidelines, vaccine administration during mild illness is considered safe and effective; current, recent, or upcoming hospitalization is not considered a contraindication to vaccination.⁴  In accordance with earlier work in other settings,^34–36  some caregivers also held misconceptions about influenza vaccine side effects, which impacted their decision-making.

Our findings indicate a need to educate families of hospitalized children. Half of respondents felt additional information would facilitate their vaccine decision-making during hospitalization. Inpatient screening at admission could help identify their specific informational needs. In this study, caregivers preferred to receive information through discussion with a doctor or nurse. Given previous data showing that 1 in 5 hospital providers did not identify influenza vaccination as a high priority for hospitalized children and felt it could cause fever and cloud the clinical picture,¹¹  it is important to assess hospital staff and provider knowledge, beliefs, and practices regarding influenza vaccination and offer influenza-specific education and vaccine communication training as needed.^37–39  Our findings provide new information regarding the importance of a strong provider recommendation for families of hospitalized children. Investigation into how and when this recommendation should be given in the hospital is needed. For families preferring to receive vaccines with their child’s primary care provider, efforts to facilitate vaccination posthospitalization are critical, especially because our data show that many remain unvaccinated.

This study has several limitations. First, the study sample was identified from a hospital-based program and represented a small subset of hospitalized families, including few from the rehabilitation and psychiatry units and a higher proportion of English-speaking, non-Hispanic white, and privately insured families, which may have biased our findings. Nonetheless, our response rate was high among families identified as eligible for the influenza survey, and demographic characteristics did not differ by survey completion status. Second, the sample of caregivers with vaccine-eligible children was small, potentially limiting power to detect associations between survey responses and influenza vaccine acceptance during hospitalization. Third, vaccine decision-making during hospitalization and end-of-season vaccination status were not documented for all families, and data regarding factors potentially impacting decision-making, such as presence of medical conditions at higher risk for influenza complications, admission diagnosis, and medical home status, were unavailable. Also, associations between demographic, clinical, and visit characteristics and individual survey responses were not assessed given our small sample and need for multiple comparisons. A deeper understanding of the multilevel factors impacting caregiver knowledge, attitudes, and practices is needed. Fourth, because the survey was administered within 72 hours of admission, the survey itself could have impacted influenza-related behaviors during hospitalization (ie, prompting families to inquire about influenza vaccination before discharge). Finally, the study population was predominantly English-speaking, non-Hispanic white, and highly educated, and the study was conducted in a state with a high nonmedical vaccine exemption rate.⁴⁰  Thus, our findings may not be generalizable to more diverse, less vaccine-hesitant populations.

This study reveals key factors impacting influenza vaccine decision-making among caregivers of hospitalized children, filling gaps in our understanding of this population and extending earlier work in nonhospital settings. This is a critical step in designing effective, evidenced-based interventions to capture vaccination opportunities during hospitalization. Alternative settings, including hospitals, are necessary for ensuring influenza vaccination coverage, especially when traditional settings may be less accessible like during the COVID-19 pandemic.⁴¹  Recent quality improvement work addressing inpatient influenza vaccination by using a multimodal approach targeting parents, nurses, and providers showed promise.^42,43  In future educational interventions, researchers should consider caregiver informational preferences and address specific concerns identified here, notably the belief that vaccination should be deferred when ill, while promoting the health benefits of influenza vaccination for high-risk patients.

We thank Kristen Feemster, Allison Fisher, Melissa Gilkey, Edgar Marcuse, Sean O’Leary, and Susan Rosenthal for their expert guidance with survey development. In addition, we are grateful to Libby Bliss, Kevin Bocek, Tony To, Sinear Sadang, Carlo Gangan, and Julie McGalliard for their assistance with survey recruitment, administration, and data management. Finally, we thank the Seattle Children’s Clinical Research Analytics team, particularly Jennifer Phillips, for helping us to obtain EHR data.