Abstract
Evidence regarding the high-dose (HD) vaccine’s relative vaccine effectiveness (rVE) and absolute benefit in reducing influenza-related hospitalizations compared to the standard-dose (SD) vaccine is warranted. We estimated the adjusted rVE and the number needed to vaccinate (NNV) of the HD vaccine compared to the SD vaccine among Clalit Health Services members aged ≥65 years. Among 418,603 and 393,125 members vaccinated in the 2022–2023 and 2023–2024 influenza seasons, the adjusted rVE was 27% (95% CI: −12% to 61%) for 2022–2023 and 7% (95% CI: −36% to 42%) for 2023–2024, with NNV to prevent one hospitalization event being 2262 (95% CI: 1004 to ∞) and 7662 (95% CI: 1293 to ∞), respectively. Even among the highest-risk subgroup, the NNV was 1289 (95% CI: 571 to ∞) for 2022–2023 and 4719 (95% CI: 797 to ∞) for 2023–2024. The HD vaccine exhibited a limited incremental benefit, even for individuals at the highest risk.
Subject terms: Influenza virus, Epidemiology, Vaccines
Introduction
Annual vaccination is the primary and most effective way to prevent influenza-associated complications, especially for high-risk groups1: people aged 65 years and above, infants, pregnant women, and people with chronic medical conditions.
There are several types of influenza vaccines: inactivated influenza vaccine, recombinant influenza vaccine, and live attenuated influenza vaccine. The pivotal high-dose (HD) vaccine study by DiazGranados, C. A. et al.2 compared it with the standard-dose (SD) vaccine in adults aged 65 years or older during the 2011–2013 influenza seasons. The relative Vaccine effectiveness (rVE) for laboratory-confirmed influenza was 24% (95% CI:10–37%). Accordingly, the US Centers for Disease Control and Prevention (CDC) recommends that patients aged 65 years or older receive enhanced influenza vaccines, including the HD vaccine, due to their higher rates of severe events and reduced immune response3. In Israel, per Ministry of Health guidelines, every person aged 6 months and above is eligible for the SD influenza vaccine4, consistent with the US CDC’s criteria1. In CHS, the HD vaccine in the study period was prioritized to very high-risk individuals 65 years and older, and additionally, those living in nursing homes, home-bound or bed-ridden patients, or patients with certain medical conditions.
However, the pivotal study was not powered to estimate the rVE for protocol-defined influenza hospitalizations, with an event rate of 6/15,990 (0.038%) in the HD group and 10/15,993 (0.063%) in the SD group. A recent study of a quadrivalent recombinant HD influenza vaccine in adults younger than 65 years5, demonstrated an rVE of 15% for lab-confirmed influenza but was not significantly more protective against influenza-related hospitalization than the SD vaccines.
The first randomized controlled trial (RCT)6 comparing the HD and SD influenza vaccines was conducted during the 2009–2010 season in adults ≥65 years. While the HD vaccine was safe, well-tolerated, and provided a superior immune response, its rVE against laboratory-confirmed influenza was not significant (12.6%, 95% CI: −140.5% to 65.8%). The authors concluded that a larger trial was needed to demonstrate superior efficacy.
Follow-up studies have yielded mixed results. One study7 found borderline significance for HD versus SD in preventing all-cause hospitalizations (rVE: 6.9%, 95% CI: 0.5–12.8%), with data from the second year contributing more significantly (rVE year 2: 13.6%, 95% CI: 5.1–21.4%). Variations in influenza activity and vaccine strain match across the two seasons may explain these discrepancies. Another trial8 in nursing home residents reported a 12% reduction in respiratory-related hospital admissions with the HD vaccine (adjusted rVE: 2–22%, p = 0.023), though the calculation of NNV was based on unadjusted data, and influenza-specific hospitalization data were not reported.
Additional studies examined HD vaccine effectiveness in high-risk populations. A pragmatic RCT9 found no significant reduction in all-cause mortality or cardiopulmonary hospitalizations, and a cluster RCT in nursing homes10 showed no significant difference in hospitalizations for major cardiovascular events. However, a recent Danish study11 during the 2021-2022 season found a 64% reduction in hospitalization for influenza or pneumonia with the HD vaccine (95% CI: 24–84%), though it did not report rVE separately for influenza hospitalizations. A recent meta-analysis study found that the rVE of the HD vaccine compared to the SD varies between 23.3–33.2%12. Yet, the evidence for HD vaccine effectiveness remains mixed, with season-to-season differences and varying influenza strain matches affecting the results13.
A decade after the introduction of the HD vaccine, there is limited evidence concerning HD’s rVE and absolute benefit compared to the SD vaccine in preventing influenza-related hospitalizations. This gap in knowledge is particularly important in the post-COVID-19 era, where the landscape of vaccine effectiveness may have changed due to shifts in viral circulation patterns and immune responses. Therefore, we aimed to estimate the rVE and number needed to vaccinate (NNV with the HD vaccine to prevent influenza-related hospitalizations, providing valuable insights for decision-making in the current context.
Results
SD vaccine population
Among the 751,895 members of CHS, 409,875 met the eligibility criteria and received the SD vaccine during the 2022–2023 influenza season, while 385,062 received it during the 2023–2024 season. The median age of the SD vaccine recipients was 74 years (IQR: 70–81 in the 2022–2023 season and 70–80 in the 2023–2024 season), with the majority being female (54.0% in the 2022–2023 season and 53.6% in the 2023–2024 season; see Supplementary Table 5).
The number of influenza hospitalizations among SD vaccine recipients was 669, with an event rate of 0.163 (95% CI: 0.151–0.176), during the 2022–2023 influenza season, and 706 cases, with an event rate of 0.183 (95% CI: 0.170–0.197), in the 2023–2024 season (Table 1).
Table 1.
Risk of influenza hospitalization following standard-dose vaccine
| Influenza season | Risk group* | N SD | Events SD | Event rate % (95% CI) ** |
|---|---|---|---|---|
| 2022–2023 | Total | 409,875 (100%) | 669 | 0.163 (0.151–0.176) |
| Low risk | 254,963 (62%) | 224 | 0.088 (0.076–0.100) | |
| Medium-high risk | 154,912 (38%) | 445 | 0.287 (0.261–0.314) | |
| 2023–2024 | Total | 385,062 (100%) | 706 | 0.183 (0.170–0.197) |
| Low risk | 233,241 (61%) | 254 | 0.109 (0.096–0.123) | |
| Medium-high risk | 151,821 (39%) | 452 | 0.298 (0.271–0.325) |
*Risk groups were determined empirically. Medium to high: (1) individuals aged 75 y and above with at least two comorbidities, (2) individuals younger than 75 y with cardiovascular background, and (3) individuals who are home-bounded.
**95% CI was based on binomial distribution.
Adjusted rVE of the HD vaccine
During the influenza season of 2022–2023, 8728 received the HD vaccine, while 8063 received it in 2023–2024 (Table 2). The HD recipients were matched with a comparable group of SD recipients, covering 403,065 and 377,126 individuals vaccinated with the SD vaccine in the 2022–2023 and 2023–2024 seasons, respectively.
Table 2.
Characteristics of the study population high-dose vaccine (HD) and matching controllers standard-dose vaccine (SD)
| 2022–2023 Influenza season | 2023–2024 Influenza season | |||
|---|---|---|---|---|
| High-dose vaccine n = 8728 (%) |
Standard-dose vaccine (n = 403,065) * Mean percentage (95% CI) |
High-dose vaccine n = 8063 (%) |
Standard-dose vaccine (n = 377,126) ** Mean percentage (95% CI) |
|
| Sex | ||||
| Female | 5925 (67.9%) | 67.9% (67.9–67.9%) | 5411 (67.1%) | 67.1% (67.1–67.1%) |
| Male | 2803 (32.1%) | 32.1% (32.1–32.1%) | 2652 (32.1%) | 32.1% (32.1–32.1%) |
| Age (years) | ||||
| Median (IQR) | 85 (77–90) | 85 (77–90) | 84 (76–89) | 85 (77–89) |
| 65–69 | 610 (7.0%) | 7.0% (7.0–7.0%) | 510 (6.3%) | 6.3% (6.3–6.3%) |
| 70–74 | 1108 (11.5%) | 11.5% (11.5–11.5%) | 880 (10.9%) | 10.9% (10.9–10.9%) |
| 75–79 | 1207 (13.8%) | 13.8% (13.8–13.8%) | 1193 (14.8%) | 14.8% (14.8–14.8%) |
| 80–84 | 1509 (17.3%) | 17.3% (17.3–17.3%) | 1332 (16.5%) | 16.5% (16.5–16.5) |
| 85+ | 4397 (50.4%) | 50.4% (50.4–50.4%) | 4150 (51.5%) | 51.5% (51.5–51.5) |
| Sociodemographic score*** | ||||
| 1–4 | 1913 (21.9%) | 26.8% (25.9–27.7%) | 1830 (22.7%) | 26.1% (25.2–27.0%) |
| 5–7 | 3536 (40.5%) | 45.0% (44.9–47.0%) | 3575 (44.3%) | 45.7% (44.6–46.8%) |
| 8–10 | 2695 (34.0%) | 24.3% (23.5–23.1%) | 2422 (30.0%) | 25.2% (24.2–26.1%) |
| Unspecified | 311 (3.6%) | 2.9% (2.6–3.3%) | 236 (2.9%) | 3.0% (2.7–3.4%) |
| Comorbidities | ||||
| Cardiometabolic | 6576 (75.3%) | 75.3% (75.3–75.3%) | 6137 (76.1%) | 76.1% (76.1–76.1%) |
| Transplant | 51 (0.6%) | 0.6% (0.6–0.6%) | 87 (1.1%) | 1.1% (1.1–1.1%) |
| Malignancy | 2260 (25.9%) | 25.9% (25.9–25.9%) | 2002 (24.8%) | 24.8% (24.8–24.8%) |
| Pulmonary | 3794 (43.5%) | 43.5% (43.5–43.5%) | 3464 (43.0%) | 43.0 (43.0–43.0%) |
| Others or no comorbidities | 1145 (13.1%) | 13.1% (13.1–13.1%) | 961 (11.9%) | 11.9% (11.9–11.9%) |
| Population sector | ||||
| General Jewish | 8480 (97.1%) | 84.7% (84.1–85.5%) | 7672 (95.1%) | 85.0% (84.3–85.8%) |
| Ultra-Orthodox Jewish | 106 (1.2%) | 3.7% (3.3–4.1%) | 255 (3.2%) | 3.4% (3.0–3.8%) |
| Arab | 145 (1.7%) | 11.6% (10.9–12.2%) | 138 (1.7%) | 11.6% (10.9–12.3%) |
| Residency in a nursing home or bedridden | ||||
| Yes | 5146 (58.9%) | 58.9% (58.9–58.9%) | 4387 (54.4%) | 54.4% (54.4–54.4%) |
| No | 3585 (41.1%) | 41.1% (41.1–41.1%) | 3678 (45.6%) | 45.6% (45.6–45.6) |
*Based on 1:1 bootstrap matching (i.e., 8728 HD patients matched with 8728 SD controls in each repetition) with 1000 repetitions. The total matching included 403,065 recipients of the SD vaccine (see Supplementary Table 5).
**Based on 1:1 bootstrap matching (i.e., 8063 HD patients matched with 8063 SD controls in each repetition) with 1000 repetitions. The total matching included 377,126 recipients of the SD vaccine (see Supplementary Table 5).
***For a detailed explanation of the sociodemographic score, please see Supplement Note 1.
The matched SD recipients exhibited similar characteristics to the HD recipients (Table 2). For instance, the majority of both the HD and SD groups were 85+ years old (50.4% in the 2022–2023 season and 51.5% in the 2023–2024 season), and two-thirds were female. Additionally, 87% of the population had comorbidities, primarily cardio-metabolic and pulmonary conditions, and 25% had malignancies. More than half of each population (58.9% in the 2022–2023 season and 54.4% in the 2023–2024 season) lived in nursing homes or were bedridden.
Among the HD recipients, the primary outcome occurred in 18 cases in 2022–2023 and in 27 cases in 2023-2024, resulting in rVE of 27% (95% CI: −12% to 61%) for 2022–2023 influenza season and 7% (95% CI: −36% to 42%) for 2023–2024 influenza season (Table 3). The combined rVE across both seasons was 17% (95% CI: −14% to 41%). Our validation survival analysis showed similar trends with a non-significant rVE (see Supplementary Note 2 and Supplementary Table 4).
Table 3.
Relative vaccine effectiveness (rVE) and NNV by risk groups
| Influenza season | Risk group* | Adjusted rVE ** (95% CI) [%] | NNV‡ (95% CI) |
|---|---|---|---|
| 2022–2023 | All | 27% (−12–61%) | 2262 (1012–) |
| Low risk | 4216 (1866–) | ||
| Medium-high risk | 1289 (571–) | ||
| 2023–2024 | All | 7% (−36% – 42%) | 7662 (1293–) |
| Low risk | 12,901 (2178–) | ||
| Medium-high risk | 4719 (797–) |
*Risk groups were determined empirically. Medium to high: (1) individuals aged 75 y and above with at least two comorbidities, (2) individuals younger than 75 y with cardiovascular background, and (3) individuals who are home-bounded.
**Based on 1:1 bootstrap matching with 1000 repetitions. The cumulative distribution of vaccination timing by vaccine type is presented in Supplementary Fig. 1.
‡The NNV was based on the computed adjusted rVE. NNV estimates should be interpreted with caution due to wide confidence intervals and an infinite upper bound (resulting from a non-significant rVE).
Absolute benefit and number needed to vaccinate
The NNV for HD compared to SD to prevent one hospitalization due to influenza was 2262 (95% CI: 1102 to ∞) during the influenza season of 2022–2023 and 7662 (1293–) in 2023–2024 (Table 3).
Subgroup analysis by baseline risk
In our post-hoc analysis, within the medium-high risk group, the primary outcome occurred in 445 cases with an event rate of 0.287 (95% CI: 0.261–0.314) during the 2022–2023 influenza season, and in 452 cases with an event rate of 0.298 (95% CI: 0.271–0.325) during the 2023–2024 influenza season (Table 1).
The NNV for HD compared SD to prevent one hospitalization due to influenza among the medium-high-risk population was determined to be 1289 (571–) in 2022–2023 and 4719 (797–) in 2023–2024 (Table 3).
Discussion
Our study demonstrates a non-significant adjusted rVE of 27% (95% CI: −12–61%) during the influenza season of 2022–2023 and 7% (95% CI: −36% to 42% in 2023–2024. The combined adjusted rVE across both seasons was 17% (95% CI: −14% to 41%). As the absolute risk for influenza hospitalization for those receiving SD is relatively low, the incremental benefit of the HD vaccine is relatively low.
Our study’s point estimate for rVE for influenza hospitalizations is similar to the rVE for lab-confirmed influenza in the pivotal RCT2–24.2% (9.7–36.5%). In our study, the absolute risk for hospitalizations due to influenza in the control group was very low (17 per 10,000 participants). However, this figure is more than double that of the control group in the pivotal RCT2, whose rate was 6 per 10,000 participants. Therefore, our study had more power to evaluate influenza hospitalizations, an outcome with higher clinical importance than lab-confirmed influenza.
Our rVE is lower than the estimated rVE in a recent pragmatic randomized feasibility trial (64%; 95% CI: 24–85%) conducted in Denmark11. However, the outcome of the Danish study was hospitalizations due to influenza or pneumonia, whereas our study evaluated hospitalizations due to influenza specifically. Therefore, this higher effect may result from pneumonia vaccinations not reported in the Danish study.
The most recent study analyzed the rVE for an HD recombinant flu vaccine (FluBlok) vs an SD flu vaccine in adults aged 50–64 years old5. Although the researchers found a very small absolute reduction in the risk of influenza with HD recombinant vaccines, there was no impact on hospitalization. The researchers reported that the HD recombinant vaccine was 15% more effective (95% CI 6.0–24.5%) in persons aged 50–64 years. However, the actual number of cases of PCR-confirmed influenza was 2.0 cases/1000 persons in the HD group and 2.34 cases/1000 persons in the SD group, which means an NNV of 3000 to prevent one more infection, with no impact on hospitalization.
Our study presents several limitations. Firstly, our analysis may not fully represent all individuals eligible to receive influenza vaccines per the guidelines outlined by the Israeli Ministry of Health, as it focuses solely on those aged 65 years and above who actually received the vaccines. Secondly, our analysis relies on data from two influenza seasons. Past evidence indicates that influenza vaccine effectiveness can vary significantly across seasons, ranging from 10% to 60%, as reported by the CDC14. Nevertheless, we emphasize that, according to the Israeli Ministry of Health report, the vaccine formula aligned with the circulating influenza virus during the 2022–2023 and 2023–2024 seasons15–18. Nonetheless, the generalizability of our findings may still be limited, as VE can fluctuate due to differences in circulating strains and vaccine matches across seasons. Thirdly, our study relied on ICD-9 codes rather than PCR confirmation to identify influenza-related hospitalizations. It is possible that some patients hospitalized with influenza-related complications were either not tested for influenza or were tested too late for detection, which could lead to outcome misclassification. Although testing decisions in hospitals generally do not differ by vaccine type (HD vs. SD), non-differential misclassification could still bias our estimates toward the null and potentially underestimate the true vaccine effect. We believe any bias introduced is likely modest given the lack of evidence suggesting systematic differences in testing practices by vaccine type.
Due to our sample size of ~8000 participants per season, we were unable to account for all potential confounders (e.g., body mass index, ethnicity, month of vaccination) or to adjust for differences in sociodemographic score and population sector. However, our primary analysis controlled for the main confounders as recommended by the CDC19.
The primary outcome measured in this study was hospitalization with a primary or secondary diagnosis of influenza, excluding hospitalizations for pneumonia due to unspecified pathogens. However, it is likely that some cases of unspecified pneumonia are complications resulting from influenza. Given that our findings indicated low or no rVE, the decision to exclude unspecified pneumonia cases actually biases the results in favor of the vaccine.
Our study explicitly considers hospitalizations as an endpoint for influenza cases. As a result, our vaccine effectiveness calculations do not account for influenza cases treated in outpatient clinics that do not require hospitalization. Lastly, it’s essential to note that the event rates in our study are low, potentially impacting the confidence interval range. However, it’s worth mentioning that these rates are higher compared to a previous study2.
There are several types of influenza vaccines: inactivated influenza vaccine, recombinant influenza vaccine, and live attenuated influenza vaccine. In Israel, vaccination is recommended for all people aged 65 years or older, infants, and high-risk groups (pregnant women, elderly population, people with chronic medical conditions, and health care workers). However, the HD vaccine is prioritized to the very high-risk patient population: 65 years and older, and living in nursing homes, or home-bound, or with chronic diseases. The cost of the HD vaccine is significantly higher than the SD vaccination. In Israel, for example, it is five times more expensive, with an additional cost of about $30. Therefore, in the study population, the corresponding cost to prevent one influenza hospitalization is estimated at $67,860 (95% CI: $30,360–$∞) for influenza season 2022–2023 and at $229,860 (95% CI: $38,790–$∞) in influenza season 2023–2024.
Our findings suggest a limited benefit of the HD influenza vaccination for preventing influenza hospitalizations in the 65+ years old population. Given the variability in vaccine effectiveness across seasons and the higher cost of the HD vaccine, healthcare policymakers should carefully assess its cost-effectiveness and consider the circumstances under which its use may be most appropriate.
The HD vaccine is associated with a non-significant lower hospitalization rate due to influenza than the SD vaccine. Moreover, due to the rarity of these events in the population vaccinated with the SD vaccine and the significantly higher cost, our results suggest that the HD vaccine should be considered with prudence.
Methods
Procedures
We analyzed the electronic medical records of individuals aged 65 years and above who are members of the largest healthcare organization in Israel, Clalit Health Services (CHS), and who received the HD or SD influenza vaccines from 1 Sept 2022 to 1 April2024 (Fig. 1, see study protocol in Supplementary Note 1; see vaccine products in Supplementary Table 1, adjuvanted influenza vaccines were not included). CHS is the largest healthcare provider in Israel, serving about 52% of the population (~4.7 million members, and two-thirds of the subjects aged 65 years or older). CHS’s members are representative of the Israeli population and reflect all demographic, ethnic, and socio-economic groups20. Records are automatically collected and updated daily in the databases of all CHS medical facilities nationwide. The study data are coded, pseudonymized, stored, and processed within the CHS research room. We extracted each patient’s socio-demographic and potential chronic illnesses information. The following demographic data were extracted for each participant: year of birth, sex, geographical district of primary healthcare clinic, population sector (General Jewish, Arab, and Ultra-orthodox Jews), and the score for socio-economic status (SES; see details in Supplement Note 1), nursing home, home-bound or bed-ridden patients, types of comorbidities (cardio-metabolic, transplant, malignancy, pulmonary). The age of each participant was calculated based on the calendar year in relation to the start of the influenza season.
Fig. 1. Study participants.
A Eligible participants for inclusion in the analyses. B Participants included in the influenza season 2022–2023 analysis. C Participants included in the influenza season 2023–2024 analysis.
Study design and participants
We conducted a retrospective, population-based comparison to assess the difference and the ratio in the risk of hospitalizations with a primary or secondary diagnosis of influenza following inoculation. The study period was the 2022–2023 influenza season, defined as 1 Sept 2022 to 1 April 2023, and the 2023–2024 influenza season, defined as 1 Sept 2023 to 1 April 2024. We also examined the risk of influenza-related hospitalizations among individuals who received the SD vaccine. To be included in our analysis, individuals needed to be alive and active members of CHS during the entire study period.
The influenza vaccination season in Israel typically takes place each year between September and February. During the 2022–2023 and 2023–2024 seasons, influenza cases in Israel began rising in November and December17, with nearly all individuals vaccinated before this increase in cases (see Supplementary Fig. 1). Additionally, the Israeli Ministry of Health’s vaccination protocol requires individuals to be free of illness at the time of vaccination16. Therefore, it is unlikely that individuals contracted influenza within at least a six-month window prior to receiving the vaccine.
Outcomes
The primary outcome measured in this study was hospitalization with a primary or secondary diagnosis of influenza (see Supplementary Table 3 for details). Our analysis also focused on assessing the risk of influenza-related hospitalization among individuals who received the SD vaccine. Additionally, we investigated the differences in risk and risk ratios for influenza hospitalizations by comparing individuals who received the HD vaccine with those who received the SD vaccine. This approach allows for a comprehensive evaluation of the effectiveness of the HD vaccine in preventing influenza-related hospitalizations compared to the SD vaccine.
Statistical analysis
Comparison between HD and SD vaccine
We used 1:1 bootstrap matching with 1000 repetitions. Specifically, in each repetition, each recipient of the HD vaccine was matched with a recipient of the SD vaccine based on variables associated with their hospitalization risk. These variables included age (categorized into five groups: 65–69 years, 70–74 years, 75–79 years, 80–84 years, and 85 years), sex (female or male), nursing home residency, bed-ridden or home-bound status (yes or no), and specific co-existing conditions outlined by the CDC as risk factors for severe influenza19: cardio-metabolic disease (yes or no), organ transplant (yes or no), malignancy (yes or no), and pulmonary disease (yes or no) (refer to Supplementary Table 2).
This approach created a possible 320 subgroups (5 × 2 × 2 × 2 × 2 × 2 × 2). However, in practice, some subgroups had no eligible HD recipients, reducing the actual number of subgroups to 197 in 2022–2023 and 210 in 2023–2024. When multiple potential SD recipients were available for a match, one was selected randomly. This method allowed us to maximize the use of SD recipients. In total, 403,065 and 377,126 unique SD recipients were selected at least once across the 1000 bootstrap repetitions for the 2022–2023 and 2023–2024 influenza seasons, respectively (Fig. 1). This approach offers increased stability and accuracy compared to commonly used methods21,22, as it produces more stable and generalizable estimates by repeatedly sampling from a large control pool.
Subsequently, we computed and adjusted the relative risk (RR) of influenza-related hospitalizations, comparing HD recipients with SD recipients. To derive the 95% confidence interval for the difference, without making assumptions about specific distributions, we employed a non-parametric percentile bootstrap method with 1000 iterations. The rVE was then calculated as 1 minus the adjusted RR.
To validate our results, we calculated the rVE using a 1:1 matching analysis with Kaplan–Meier estimates (see Supplementary Note 2). This analysis was performed under similar assumptions to our primary analysis, with two key differences. First, to improve the accuracy of matching, we also matched individuals based on the number of previous influenza vaccinations received over specified years (0, 1–3, 4+) and any hospitalizations in the prior year (0, 1, 2, 3+). This approach aligns with a large-scale study demonstrating that vaccination compliance and healthcare usage are associated with an individual’s risk of complications, thereby adding a behavioral layer to reduce potential biases23. Second, for individuals who died during the influenza season, we censored data for each matched pair at the time of death of either member, consistent with a previous study24.
In further analysis, we determined the NNV with the HD vaccine to prevent one hospitalization among the recipients of the SD vaccine. This was based on the absolute risk reduction, assuming the computed rVE25.
We repeated the aforementioned analysis for each season individually. Additionally, we calculated the overall rVE by combining data from both seasons, assuming independence between them.
Subgroup analysis
As post-hoc analysis, we classified recipients of the SD vaccine into two risk categories in a data-driven approach: low and medium to high, based on their susceptibility to hospitalization due to influenza following inoculation. First, we categorized the individuals into subpopulations based on several factors: age (grouped as 65–69 years, 70–74 years, 75–79 years, 80–84 years, and ≥85 years), sex (female or male), nursing home residency, bed-ridden or home-bound status (yes or no), cardiovascular conditions (yes or no), and the number of co-existing conditions other than cardiovascular conditions (1, 2, or 3+). Next, we calculated the risk of influenza-related hospitalizations for each subpopulation. Finally, we identified the “medium to high-risk” group by selecting subpopulations that fall within the 18th percentile based on their influenza-related hospitalization risk.
For the “medium to high-risk” group, we identified (1) individuals aged 75 years and above with at least two comorbidities, (2) individuals younger than 75 years with a cardiovascular background, and (3) individuals who are home-bounded. All other recipients fell into the “low-risk” category.
For each level of risk, we calculated the risk of hospitalization along with its 95% confidence interval, assuming a binomial distribution. Consequently, we determined the NNV with high dose vaccine to prevent one hospitalization for each risk level based on the absolute risk reduction assuming the computed rVE.
Ethical approval
As the retrospective data was pseudonymized, the institutional Helsinki review board and data utilization committee approved using the retrospective cohort data without requiring specific consent from the members of Clalit Health Services (protocol number 0139-21-CHS).
Supplementary information
Acknowledgements
This research was partially supported by the European Research Council (ERC) project #949850 and the Israel Science Foundation (ISF), grant No. 3409/19, within the Israel Precision Medicine Partnership program. The sponsor of the study had no role in the study's design, data collection, data analysis, and data interpretation or in the writing of the report. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.
Author contributions
Conception and design: D.Y., M.Y., E.S., R.A., and S.Y. Collection and assembly of data: T.R. and M.Y. Analysis and interpretation of the data: M.Y., D.Y., E.S., and R.A. Statistical expertise: M.Y., D.Y., E.S. Drafting the article: M.Y., R.A., and S.Y. Critical revision of the article for important intellectual content: S.Y., M.Y., D.Y., I.B., E.S., R.A., and D.N. Final approval of the article: all authors. Obtaining funding: D.Y. and E.S.
Data availability
The datasets generated and/or analyzed during the current study are not publicly available due to restrictions set by the CHS Helsinki and data utilization committees, which prohibit sharing patient-level data outside the permitted researchers. However, the aggregated data are available from the corresponding author upon reasonable request.
Code availability
The underlying statistical code for this study is not publicly available but may be made available to qualified researchers on reasonable request from the corresponding author.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
These authors contributed equally: Shlomit Yaron, Matan Yechezkel, Dan Yamin.
These authors jointly supervised this work: Ronen Arbel, Doron Netzer
Supplementary information
The online version contains supplementary material available at 10.1038/s41541-025-01065-5.
References
- 1.Who needs a flu vaccine | CDC. https://www.cdc.gov/flu/prevent/vaccinations.htm (2024).
- 2.DiazGranados, C. A. et al. Efficacy of high-dose versus standard-dose influenza vaccine in older adults. N. Engl. J. Med.371, 635–645 (2014). [DOI] [PubMed] [Google Scholar]
- 3.Fluzone high-dose seasonal influenza vaccine | CDC. https://www.cdc.gov/flu/vaccine-types/flu-zone.html (2024).
- 4.Influenza | Ministry of Health. https://www.gov.il/en/Departments/Guides/disease-flu?chapterIndex=2.
- 5.Hsiao, A. et al. Recombinant or standard-dose influenza vaccine in adults under 65 years of age. N. Engl. J. Med.389, 2245–2255 (2023). [DOI] [PubMed] [Google Scholar]
- 6.DiazGranados, C. A. et al. High-dose trivalent influenza vaccine compared to standard dose vaccine in elderly adults: safety, immunogenicity and relative efficacy during the 2009-2010 season. Vaccine31, 861–866 (2013). [DOI] [PubMed] [Google Scholar]
- 7.DiazGranados, C. A. et al. Prevention of serious events in adults 65 years of age or older: a comparison between high-dose and standard-dose inactivated influenza vaccines. Vaccine33, 4988–4993 (2015). [DOI] [PubMed] [Google Scholar]
- 8.Gravenstein, S. et al. Comparative effectiveness of high-dose versus standard-dose influenza vaccination on numbers of US nursing home residents admitted to hospital: a cluster-randomised trial. Lancet Respir. Med.5, 738–746 (2017). [DOI] [PubMed] [Google Scholar]
- 9.Vardeny, O. et al. Effect of high-dose trivalent vs standard-dose quadrivalent influenza vaccine on mortality or cardiopulmonary hospitalization in patients with high-risk cardiovascular disease: a randomized clinical trial. JAMA325, 39–49 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Saade, E. A. et al. High-dose influenza vaccines for the prevention of hospitalization due to cardiovascular events in older adults in the nursing home: post-hoc analysis of a cluster-randomized trial. Vaccine40, 6700–6705 (2022). [DOI] [PubMed] [Google Scholar]
- 11.Johansen, N. D. et al. A pragmatic randomized feasibility trial of influenza vaccines. NEJM Evid.2, EVIDoa2200206 (2023). [DOI] [PubMed] [Google Scholar]
- 12.Skaarup, K. G. et al. The relative vaccine effectiveness of high-dose vs standard-dose influenza vaccines in preventing hospitalization and mortality: a meta-analysis of evidence from randomized trials. J. Infect.89, 106187 (2024). [DOI] [PubMed] [Google Scholar]
- 13.Lu, Y. et al. Effect of age on relative effectiveness of high-dose versus standard-dose influenza vaccines among US medicare beneficiaries aged ≥65 years. J. Infect. Dis.220, 1511–1520 (2019). [DOI] [PubMed] [Google Scholar]
- 14.Past Seasons’ Vaccine Effectiveness Estimates | CDC. https://www.cdc.gov/flu/vaccines-work/past-seasons-estimates.html (2024).
- 15.ICDC–Surveillance report on morbidity due to respiratory viruses season 2022–2023 (Hebrew). https://www.gov.il/BlobFolder/reports/flu-2022–2023/he/files_weekly-flu-corona_flu_he_flu-2022–2023.pdf (2023).
- 16.Unit of epidemiology Israeli ministry of health: Influenza vaccine guidelines 2022. https://www.health.gov.il/unitsoffice/hd/ph/epidemiology/td/docs/360_influenza.pdf (2022).
- 17.Surveillance report on morbidity due to respiratory viruses week 17 ending 27/4/24. https://www.gov.il/BlobFolder/reports/flu-27042024/en/files_weekly-flu-corona_flu_en_flu-27042024e.pdf (2024).
- 18.Recommended composition of influenza virus vaccines for use in the 2023–2024 northern hemisphere influenza season. https://www.who.int/publications/m/item/recommended-composition-of-influenza-virus-vaccines-for-use-in-the-2023–2024-northern-hemisphere-influenza-season (2023).
- 19.https://www.cdc.gov/flu/highrisk/index.htm People at Increased Risk for Flu Complications | Influenza (Flu) | CDC (2024).
- 20.Cohen R., Rabin H. National Insurance Institute of Israel. Membership in sick funds 2016. (In Hebrew). https://www.btl.gov.il/Publications/survey/Documents/seker289/seker_289.pdf (2016).
- 21.Austin, P. C. & Small, D. S. The use of bootstrapping when using propensity-score matching without replacement: a simulation study. Stat. Med.33, 4306 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Hill, J. & Reiter, J. P. Interval estimation for treatment effects using propensity score matching. Stat. Med.25, 2230–2256 (2006). [DOI] [PubMed] [Google Scholar]
- 23.Shaham, A., Chodick, G., Shalev, V. & Yamin, D. Personal and social patterns predict influenza vaccination decision. BMC Public Health20, 222 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Barda, N. et al. Safety of the BNT162b2 mRNA Covid-19 vaccine in a nationwide setting. N. Engl. J. Med.385, 1078–1090 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Arbel, R. et al. Effectiveness of a bivalent mRNA vaccine booster dose to prevent severe COVID-19 outcomes: a retrospective cohort study. Lancet Infect. Dis.23, 914–921 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The datasets generated and/or analyzed during the current study are not publicly available due to restrictions set by the CHS Helsinki and data utilization committees, which prohibit sharing patient-level data outside the permitted researchers. However, the aggregated data are available from the corresponding author upon reasonable request.
The underlying statistical code for this study is not publicly available but may be made available to qualified researchers on reasonable request from the corresponding author.