ABSTRACT
As higher-valent pneumococcal conjugate vaccines (PCVs) become available for pediatric populations in the US, it is important to understand healthcare provider (HCP) preferences for and acceptability of PCVs. US HCPs (pediatricians, family medicine physicians and advanced practitioners) completed an online, cross-sectional survey between March and April 2023. HCPs were eligible if they recommended or prescribed vaccines to children age <24 months, spent ≥25% of their time in direct patient care, and had ≥2 y of experience in their profession. The survey included a discrete choice experiment (DCE) in which HCPs selected preferred options from different hypothetical vaccine profiles with systematic variation in the levels of five attributes. Relative attribute importance was quantified. Among 548 HCP respondents, the median age was 43.2 y, and the majority were male (57.9%) and practiced in urban areas (69.7%). DCE results showed that attributes with the greatest impact on HCP decision-making were 1) immune response for the shared serotypes covered by PCV13 (31.4%), 2) percent of invasive pneumococcal disease (IPD) covered by vaccine serotypes (21.3%), 3) acute otitis media (AOM) label indication (20.3%), 4) effectiveness against serotype 3 (17.6%), and 5) number of serotypes in the vaccine (9.5%). Among US HCPs, the most important attribute of PCVs was comparability of immune response for PCV13 shared serotypes, while the number of serotypes was least important. Findings suggest new PCVs eliciting high immune responses for serotypes that contribute substantially to IPD burden and maintaining immunogenicity against serotypes in existing PCVs are preferred by HCPs.
Introduction
Pneumococcal disease is caused by the gram-positive bacterium Streptococcus pneumoniae and can be classified as invasive, such as bacteremic pneumonia, bacteremia without focus, and meningitis, or noninvasive, such as non-bacteremic pneumococcal pneumonia and acute otitis media (AOM).Citation1,Citation2 The impact of pneumococcal conjugate vaccines (PCVs) in the United States (US) was reflected by the significant decrease in pneumococcal disease. From 1998 to 2019, there was a 93% reduction in prevalence among children <5 y old.Citation3 Despite this reduction, certain vaccine-type serotypes, specifically 3, 19A, and 19F, continue to disproportionately contribute to disease burden in the US and S. pneumoniae is estimated to cause more than 300,000 deaths in children under the age of five globally each year.Citation4,Citation5
There are various pneumococcal vaccines that have been approved for use in pediatric populations in the US for the prevention of invasive pneumococcal disease. In 2000, the US Food and Drug Administration (FDA) approved a 7-valent PCV (PCV7) for the prevention of invasive pneumococcal disease (IPD) in infants aged <24 months and otitis media in infants and toddlers. In 2010, the FDA approved a 13-valent PCV (PCV13) for the prevention of IPD and otitis media (the otitis media indication covers the 7 serotypes included in PCV7) in children 6 weeks to 5 y old and the approval was expanded in January 2013 to include the prevention of IPD in children 6 to 17 y old.Citation6,Citation7 Most recently, in June 2022, PCV15 was approved for the prevention of IPD in infants and children 6 weeks through 17 y old and in April 2023,Citation8 PCV20 was approved for the prevention of IPD in infants and children 6 weeks through 17 y old and for the prevention of otitis media (caused by the 7 serotypes included in PCV7) in infants 6 weeks through 5 y old.Citation9
Previously, the Advisory Committee on Immunization Practices (ACIP) recommended PCV13 as the standard of care for US children, however, in June 2022, recommendations were amended to include both PCV13 and PCV15 and were amended again in June 2023 to include both PCV15 and PCV20 as vaccine options for children.Citation8,Citation10 For children <24 months, PCV15 and PCV20 are recommended for routine use according to the existing approved schedules.Citation10
With multiple vaccine options available for use and the evolving prevention landscape in the US, it is increasingly important to understand health care providers’ (HCPs) knowledge, attitudes, and preferences regarding pneumococcal vaccination, especially considering that receiving information and a recommendation from HCPs has been associated with vaccine uptake.Citation11,Citation12 While there is existing literature on HCPs’ attitudes and preferences for pneumococcal vaccination in adults,Citation13,Citation14 little is known on HCPs’ attitudes and preferences for pneumococcal vaccination in children. The objective of this study was to assess HCPs preferences, and knowledge, attitudes, and behaviors (KAB) toward PCVs in children age <24 months in the United States to support the evaluation of the feasibility and acceptability of novel PCVs.
Materials and methods
Study design and eligibility criteria
A cross-sectional survey that included a discrete choice experiment (DCE), was conducted among HCPs (pediatricians, family medicine physicians, and advanced practitioners [nurse practitioners and physician assistants]) in the US between March 2023 and April 2023. A DCE is a research method used to quantify stakeholders’ preferences and understand key factors driving their decisions by systematically eliciting choices between hypothetical options (i.e., vaccine profiles).Citation15 All HCPs were recruited by a recruitment vendor using its existing online physician panels. A random selection of potentially eligible HCPs (potentially eligible based on physician/clinician type, but may or may not meet all the study’s inclusion and exclusion criteria) were sent invitations and those HCPs who expressed interest in participating were screened for eligibility until the target sample size of eligible participants was reached. The sample size was based on power calculations for the DCE (which indicated that ~100 respondents would be needed for each provider type to allow for possible subgroup analysis). Thus, a target of 500 respondents across all HCP types was deemed sufficient to allow for some respondent loss due to failures of comprehension checks and still achieve the target sample size. To minimize bias, HCPs were blinded to the study sponsor. HCPs were paid an honorarium for their time that was determined to be within a fair market value range, considering time commitment and study type. The study was reviewed by an institutional review board (IRB) and was deemed exempt from IRB oversight.
HCPs were eligible if they 1) provided direct patient care to pediatric patients and recommend or administered vaccines as part of patient care, 2) worked in a practice that provided vaccines to infants and children aged ≤24 months, 3) spent at least 25% of their time providing care to pediatric patients, 4) practiced within the US (50 states and Washington DC), and 5) were able to complete the survey in English. HCPs were excluded from participating if they were unwilling to provide electronic agreement to participate in the online survey or had <2 y of experience in their profession.
Survey development
Qualitative, semi-structured exploratory interviews with HCPs who provided care to pediatric patients were conducted prior to survey development to identify possible vaccine attributes and concepts for inclusion in the DCE.Citation16 Findings from a targeted literature review, exploratory interviews, and pretest interviews were used to inform the content of the survey. The survey included six sections: screener, electronic agreement form, demographic and practice characteristics, DCE attribute descriptions and comprehension checks, DCE choice tasks, and a KAB survey section. HCPs were also provided with a brief background on pneumococcal disease epidemiology. Comprehension check questions were constructed as simple exercises focusing exclusively on one attribute at a time to assess respondents’ comprehension of each attribute, such as by asking HCPs to select the level with the greatest number of serotypes. Two or more comprehension check failures, that is selecting a level with a lower number of serotypes or lower immune response as better than one with more serotypes or higher immune response, were an indicator of poor comprehension of the attributes and HCPs with two or more failures were disqualified from participating in the study.
A total of five vaccine attributes were examined in the DCE survey component: number of serotypes included in the vaccine, effectiveness against serotype 3, comparability of immune response for the serotypes shared with PCV13, percent coverage of IPD, and AOM label indication (). The combinations of attributes and levels that were presented to HCPs in the choice tasks were generated from the D-efficient Bayesian design using the Ngene software (version 1.2.1, by ChoiceMetrics). HCPs were instructed to assume similar cost and safety while reviewing hypothetical vaccine profiles. HCPs were presented with a total of 12 choice tasks, 10 of which were included in the analysis. shows an example of a possible choice task, in which each attribute is populated with a level and the participant would select the preferred hypothetical vaccine. The levels for each attribute were populated based on the survey design software and combinations were not forced; choice tasks may have resembled marketed products but were not designed specifically to do so. Of the other two, one was a repeated question to explore consistency, and the other was a dominance check with one choice superior on all attributes that served as a quality control measure. Failing the repeated question or dominance task did not automatically eliminate the HCPs from the preference analysis since the random utility model assumes an error component to account for seemingly irrational behavior.
Figure 1. Example of choice task in discrete choice experiment.
Table 1. Attributes and levels presented in the discrete choice experiment (DCE).
Data collection and study measures
Survey links were distributed via e-mail to verified HCP panelists who met screening criteria. HCPs interested in participating in the survey answered a series of questions via self-report to confirm eligibility before the start of the survey, complementing the panel’s enrollment and ongoing validation processes. Data collected included HCP demographics and practice characteristics, DCE choice task selections, and KAB regarding currently available PCVs, new PCVs, immunogenicity creep, and breakthrough disease. Immunogenicity creep was defined as the potential reduction in immune response as more serotypes are added to expanded valency PCVsCitation17 and breakthrough disease was defined as “disease caused by vaccine-type serotypes following partial vaccination.”
Data analysis
HCP demographics and practice characteristics were analyzed descriptively using univariate statistics. Mean, standard deviation, median, and range are presented for continuous variables while frequency and percent are presented for categorical variables.
Dummy-coded conditional logit regression models were used to examine the observed choice task responses in the form of odds ratios (ORs). Relative attribute importance (RAI) scores, the proportion of total variance explained by individual attributes expressed as a percentage, were computed to understand the importance HCPs place on each of the attributes. The higher the score, the greater the influence of that attribute on decision-making. Sensitivity analysis was conducted to compare results for the study sample with and without failures (i.e., excluding respondents who failed one comprehension check or the dominance question). Data analyses were conducted using Stata® 14.2 (for the analysis of the DCE) and SAS/STAT® 14.3 (for the remainder of analyses of survey data).
Results
Demographic and practice characteristics of HCPs
A total of 32 respondents failed 2 or more comprehension checks, and thus, were disqualified from participating in the study. A total of 548 remaining HCPs (332 physicians and 216 advanced practitioners) met the inclusion criteria and completed the survey. HCPs were an average of 43.2 y old, and a majority were male (57.9%), White (77.4%), and practiced in an urban location (69.7%). Hospital-owned practices were the most common settings (38.7%) and most practices/institutions had a program to incentivize childhood vaccination (53.5%). HCPs reported that more than half of their patients (50.8%) are covered under public insurance and are without chronic medical conditions (50.9%) ().
Table 2. HCP and practice characteristics.
Knowledge, attitudes, and behaviors of HCPs regarding PCVs
Most HCPs (92.3%) agreed or strongly agreed that they would like to have more approved options for patients <24 months old, with the most common reasons being flexibility to select a new PCV option based on immune responses (58.1%) and number of serotypes covered (52.8%). Most HCPs reported that they were familiar (extremely [23.5%], moderately [46.4%), or somewhat [24.6%]) with higher-valent PCVs that are newly approved or in development for pediatric patients <24 months old ().
Table 3. Knowledge, attitudes, and behaviors of HCPs regarding PCVs.
Among different information sources used in deciding to recommend a PCV, CDC recommendations (18.2%), information from the manufacturer (12.6%), and published clinical studies (12.5%) were considered the most important in the decision-making process ().
Figure 2. Most important information source in making PCV recommendations (n = 548).
Approximately half of the HCPs (50.9%) were already familiar with the concept of immunogenicity creep. Moreover, 62.0% of HCPs reported being at least somewhat concerned about immunogenicity creep (extremely [8.9%], moderately [22.1%], and somewhat [31.0%]) and 64.8% of HCPs were at least somewhat concerned for breakthrough disease with new higher-valent PCVs (extremely [9.1%], moderately [21.7%], and somewhat [33.9%]) ().
DCE choice task preferences of HCPs
Analyses were conducted for all respondents and excluding respondents who failed a single comprehension check (n = 34) or dominance question (n = 28). There were no statistical differences, so all respondents were retained in the analysis. RAI was highest for immune response for the shared serotypes covered in PCV13 (RAI score: 31.4%), followed by percent coverage of IPD (21.3%), AOM indication (20.3%), effectiveness against serotype 3 (17.6%), and number of serotypes included (9.5%) ().
Figure 3. Relative attribute importance (RAI) of PCV attributes estimated from regression models (n = 548).
HCPs were three times (OR: 3.0, 95% confidence interval [CI]: 2.5–3.6) as likely to prefer a vaccine with an immune response comparable to PCV13 on all 13 of 13 serotypes with 1 superior serotype, compared to a vaccine with an immune response comparable to PCV13 on 8 of 13 serotypes and inferior on the remaining 5 serotypes. They were also 2.1 (95% CI: 1.8–2.4) times as likely to prefer 53% coverage of IPD compared to 17% coverage and 2.0 (95% CI: 1.8–2.3) times as likely to prefer a vaccine with an AOM label indication based on data from new clinical trials compared to a vaccine with no AOM label indication. HCPs were also 1.4 (95% CI: 1.3–1.5) times as likely to prefer a PCV vaccine that covers 20 serotypes compared to a vaccine that covers 13 serotypes and 1.9 (95% CI: 1.6–2.2) times as likely to prefer a vaccine with 95% effectiveness against serotype 3 compared to 15% effectiveness (). HCPs’ preferences for effectiveness behaved in a stair-step pattern, with a trend for stronger preferences for high levels of effectiveness, although confidence intervals often overlapped.
Table 4. Odds ratio preference results estimated from regression models (n = 548).
Discussion
To our knowledge, this is the first DCE survey study to assess the knowledge, attitudes, behaviors, and preferences of US HCPs regarding PCVs in children aged <24 months. Results demonstrated that almost all HCPs were interested in having more approved PCV options for younger patients (<24 months). Most HCPs also reported concern about immunogenicity creep and breakthrough disease with newer higher-valent PCVs. Moreover, CDC recommendations were the most important information source influencing HCP decision-making regarding PCV recommendations.
In the DCE, immune response for the shared PCV13 serotypes was considered the most important factor in HCPs’ decision-making for PCVs, while the number of serotypes was considered least important. The relatively lower importance for the number of serotypes compared with attributes describing immune response and disease burden suggests that respondents are aware that serotypes do not equally contribute to the burden of disease. Results demonstrate that most HCPs in this study believe that maintaining comparable immune response for PCV13 vaccine-type serotypes, which are estimated to cause 37.4% of global IPD burden in children <5 y old,Citation18 is more important than expanding the number of serotypes.
Within each individual attribute included in the DCE choice tasks, there was a clear and expected pattern in level preferences, even though there were not always significant differences between adjacent levels. Overall, HCPs’ responses were logical and indicated that when they considered the levels of each attribute individually, they preferred a stronger immune response, higher effectiveness against serotype 3, a higher percentage of coverage against IPD, an AOM indication, and a higher number of included serotypes. It is important to note that there was no significant difference in preferences for the source of evidence (new clinical trial data vs. immunobridging data from PCV7) used for an AOM label indication, however, our findings indicate that having an AOM indication is significantly valuable to HCPs, which may be due to experience with AOM as the most common infectious disease in children <24 months old.Citation19
While there are no other published preference studies among US HCPs regarding PCVs in children <24 months old, studies related to other vaccines can provide some insights on vaccine preferences, although vaccine attributes are not defined identically. A multi-country preference study of vaccine-related attitudes and concerns of HCPs who provide pediatric care found that severity of the relevant disease and inclusion of the vaccine in formal vaccination schedules were the most important factors affecting HCPs pediatric vaccine recommendations.Citation20 Another study in China on HCPs’ preferences regarding COVID-19 vaccines for pediatric patients found that HCPs considered vaccine efficacy to be the most influential attribute on their preferences, when compared to vaccine safety and number of doses administered.Citation21 Moreover, a study on US pediatricians’ preferences for hypothetical meningococcal vaccines for infants found increases in vaccine effectiveness to be one of the most important factors influencing their recommendation.Citation22 Insights from other non-PCV specific preference studies corroborate the findings of our study that coverage of more severe disease (such as IPD) and vaccine effectiveness may heavily influence HCP decision-making regarding pediatric vaccinations.
This study was subject to limitations. First, HCPs were recruited from online panels and HCPs who participate in online panels may differ from those who do not, potentially reducing the generalizability of our results. Second, the HCP sample was predominantly White (77%), male (58%), and worked in an urban location (70%), which may not be representative of the US pediatric HCP population (less than 40% male according to the American Board of Pediatrics), nor are they likely to be representative of HCP populations outside the US. Social desirability bias may also have affected results; if HCPs were uncomfortable self-reporting limited knowledge of PCVs, they may falsely report higher knowledge or familiarity with PCVs. However, this bias was mitigated by using an online survey completed independently and confidentially. Third, the DCE required that respondents evaluated hypothetical profiles and assumed comparable safety and cost when comparing vaccine profiles, all of which may or may not reflect real-world vaccine profiles. However, qualitative work and pre-testing were undertaken to ensure that the vaccine profiles were clinically relevant and reflected characteristics of vaccines currently available and in development. Lastly, subgroup analyses were not conducted to explore whether there were different preferences for vaccine attributes across subpopulations of HCPs, which could serve as effect modifiers for our results.
This study also has several strengths. First, results help to fill an important knowledge gap about the importance of characteristics of new PCVs to HCPs as new PCVs are being introduced and incorporated into pediatric pneumococcal vaccine recommendations. While there are existing preference studies on other vaccines, none have specifically considered attributes of pediatric PCVs. Through a literature search and a qualitative phase in which we interviewed all four types of HCPs included in this study (pediatricians, family medicine physicians, nurse practitioners, and physician assistants), we developed a set of attributes and levels that were clinically relevant and can inform discussions about preferences and acceptability of newer PCVs.
Conclusion
This study found that comparability of immune response for shared serotypes covered in PCV13 and percentage coverage of IPD were the most important factors affecting HCPs’ decision to recommend a PCV, while the number of serotypes was deemed to be least important among the attributes explored. These findings suggest new PCVs eliciting high immune responses for serotypes that contribute substantially to IPD burden and maintaining immunogenicity against serotypes in existing PCVs, are preferred by HCPs.
Acknowledgments
The authors would like to thank all HCPs who participated in this study, as well as Empanel for its assistance in recruiting participants for this study. Data analysis support provided by Debdeep Chattopadhyay from OPEN Health, Rotterdam, the Netherlands.
Disclosure statement
S.M., J.W., K.A.F., B.C., and T.W.W. are employees of Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA, and may own stock/stock options in Merck & Co., Inc., Rahway, NJ, USA. Inc. J.T, N.N, A.P, R.P.V, M.H and J.K.S are employees of OPEN Health, which received consulting fees from Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA.
Additional information
Funding
References
- Scelfo C, Menzella F, Fontana M, Ghidoni G, Galeone C, Facciolongo NC. Pneumonia and invasive pneumococcal diseases: the role of pneumococcal conjugate vaccine in the era of multi-drug resistance. Vaccines (Basel). 2021 Apr 22;9(5):420. doi:10.3390/vaccines9050420.
- Tan TQ. Pediatric invasive pneumococcal disease in the United States in the era of pneumococcal conjugate vaccines. Clin Microbiol Rev. 2012 Jul;25(3):409–9. doi:10.1128/CMR.00018-12.
- Centers for Disease Control and Prevention. Surveillance and reporting; 2020 Sep 1 [accessed 2023 Jan 5]. https://www.cdc.gov/pneumococcal/surveillance.html.
- Centers for Disease Control and Prevention. Global Pneumococcal disease and vaccination; 2022 Jan 27 [accessed 2023 Jan 5]. https://www.cdc.gov/pneumococcal/global.html.
- Gierke R. Current epidemiology of pediatric pneumococcal disease, United States. https://www.cdc.gov/vaccines/acip/meetings/downloads/slides-2023-02/slides-02-22/Pneumococcal-02-Gierke-508.pdf.
- Pfizer. Pfizer receives FDA approval for prevnar 13™ for the prevention of invasive pneumococcal disease in infants and young children; 2010 Feb 24 [accessed 2023 Jan 5]. https://www.pfizer.com/news/press-release/press-release-detail/pfizer_receives_fda_approval_for_prevnar_13_for_the_prevention_of_invasive_pneumococcal_disease_in_infants_and_young_children.
- Centers for Disease Control and Prevention (CDC). Use of 13-valent pneumococcal conjugate vaccine and 23-valent pneumococcal polysaccharide vaccine among children aged 6-18 years with immunocompromising conditions: recommendations of the advisory committee on immunization practices (ACIP). MMWR Morb Mortal Wkly Rep. 2013;62(25):521–524.
- Kobayashi M, Farrar JL, Gierke R, Leidner AJ, Campos-Outcalt D, Morgan RL, Long SS, Poehling KA, Cohen AL, Poehling KA, et al. Use of 15-Valent Pneumococcal conjugate vaccine among U.S. children: updated recommendations of the advisory committee on immunization practices — United States, 2022. MMWR Morb Mortal Wkly Rep. 2022;71:1174–1181. doi:10.15585/mmwr.mm7137a3.
- Pfizer. U.S. FDA approves PREVNAR 20®, Pfizer’s 20-valent pneumococcal conjugate vaccine for infants and children; 2023 Apr 27 [accessed 2023 June 7]. https://www.pfizer.com/news/press-release/press-release-detail/us-fda-approves-prevnar-20r-pfizers-20-valent-pneumococcal.
- Centers for Disease Control and Prevention. Advisory committee on immunization practices (ACIP); 2022 Jan 28 [accessed 2023 Jan 5]. https://www.cdc.gov/vaccines/acip/index.html.
- Freed GL, Clark SJ, Butchart AT, Singer DC, Davis MM. Sources and perceived credibility of vaccine-safety information for parents. Pediatrics. 2011;127(Suppl 1):107–112. doi:10.1542/peds.2010-1722P.
- Gust DA, Darling N, Kennedy A, Schwartz B. Parents with doubts about vaccines: which vaccines and reasons why. Pediatrics. 2008;122(4):718–725. doi:10.1542/peds.2007-0538.
- Vietri J, Meyers K, Poulos C, Chilson E, Sweeney C, Davis K, Snow V. United States healthcare provider preferences for adult pneumococcal vaccine recommendations. Open Forum Infect Dis. 2021 Dec 4;8(Suppl 1):S127. doi:10.1093/ofid/ofab466.209.
- Sacco P, Myers K, Poulos C, Sweeney C, Hollis K, Snow V, Vietri JT. Preferences for adult Pneumococcal vaccine recommendations among United States health care providers. Infect Dis Ther. 2019;8(4):657–670. doi:10.1007/s40121-019-00266-5.
- Ryan M, Farrar S. Using conjoint analysis to elicit preferences for health care. BMJ. 2000;320(7248):1530–1533. doi:10.1136/bmj.320.7248.1530.
- Mohanty S, Tsai J-H, Ning N, Pena-Molina A, Verma RP, Heisen M, Weaver J, Feemster KA, Weiss T, Schmier J, et al. 583. Preferences and attitudes of healthcare providers towards pneumococcal conjugate vaccines (PCVs) for children ages two and under in the United States (US). Open Forum Infect Dis. 2022;9(Supplement_2). doi:10.1093/ofid/ofac492.635.
- Kwambana-Adams BA, Mulholland EK, Satzke C; ISPPD Group. State-of-the-art in the pneumococcal field: proceedings of the 11th international symposium on pneumococci and pneumococcal diseases (ISPPD-11). Pneumonia (Nathan). 2020 Feb 5;12:2. doi:10.1186/s41479-019-0064-y.
- Grant LR, Slack MPE, Theilacker C, Vojicic J, Dion S, Reinert R-R, Jodar L, Gessner BD. Distribution of serotypes causing invasive pneumococcal disease in children from high-income countries and the impact of pediatric pneumococcal vaccination. Clin Infect Dis. 2023;76(3):1062–1070. doi:10.1093/cid/ciac475.
- Jamal A, Alsabea A, Tarakmeh M, Safar A. Etiology, diagnosis, complications, and management of acute otitis media in children. Cureus. 2022 Aug 15;14(8):e28019. doi:10.7759/cureus.28019.
- Bakhache P, Rodrigo C, Davie S, Ahuja A, Sudovar B, Crudup T, Rose M. Health care providers’ and parents’ attitudes toward administration of new infant vaccines—a multinational survey. Eur J Pediatr. 2013;172(4):485–492. doi:10.1007/s00431-012-1904-4.
- Hou Z, Song K, Wang Q, Zang S, Tu S, Chantler T, Larson HJ. Childhood COVID-19 vaccine acceptance and preference from caregivers and healthcare workers in China: a survey experiment. Prev Med. 2022;161:107138. doi:10.1016/j.ypmed.2022.107138.
- Poulos C, Reed Johnson F, Krishnarajah G, Anonychuk A, Misurski D. Pediatricians’ preferences for infant meningococcal vaccination. Value Health. 2015;18(1):67–77. doi:10.1016/j.jval.2014.10.010.