A clinical and economic assessment of adjuvanted trivalent versus standard egg-derived quadrivalent influenza vaccines among older adults in the United States during the 2018-19 and 2019-20 influenza seasons

ABSTRACT

Background

Clinical evidence supports use of enhanced influenza vaccines in older adults. Few economic outcome studies have compared adjuvanted trivalent inactivated (aIIV3) and standard egg-derived quadrivalent inactivated influenza vaccines (IIV4e).

Research Design and Methods

A retrospective cohort study was conducted leveraging deidentified US hospital data linked to claims data during the 2018–19 and 2019–20 influenza seasons. Relative vaccine effectiveness (rVE) was compared in adults aged ≥ 65 years receiving aIIV3 or IIV4e using inverse probability of treatment weighting (IPTW) and Poisson regression. An economic assessment quantified potential real-world cost savings.

Results

The study included 715,807 aIIV3 and 320,991 IIV4e recipients in the 2018–19 and 844,169 aIIV3 and 306,270 IIV4e recipients in the 2019–20 influenza seasons. aIIV3 was significantly more effective than IIV4e in preventing cardiorespiratory disease (2018–19 rVE = 6.2%; and 2019–20 rVE = 6.0%) and respiratory disease (2018–19 rVE = 8.9%; and 2019–20 rVE = 10.1%). During the 2018–19 influenza season cardiorespiratory hospitalization cost savings for the aIIV3 population were $392 M, and $221 M for the 2019–20 season. Respiratory hospitalization cost savings for the aIIV3 population were $145 M and $97 M, respectively.

Conclusions

Our findings suggest that aIIV3 provides clinical and economic advantages versus IIV4e in the elderly.

Plain Language Summary

Flu vaccines do not work as well in older adults due to the aging of their immune system. One approach to improving vaccine efficacy is the addition of a substance, or adjuvant, to the vaccine in order to boost an individual’s immune response. This study evaluated an adjuvanted vaccine compared to an unadjuvanted vaccine for preventing cardiorespiratory hospitalizations and hospitalization costs. The findings demonstrated that the adjuvanted flu vaccine, compared to the unadjuvanted vaccine, prevented more hospitalizations and greatly reduced associated hospital costs.

KEYWORDS:

• Adjuvanted
• aIIV3
• influenza vaccine
• economic assessment
• elderly
• relative vaccine effectiveness

1. Introduction

Older adults (≥65 years of age) have a greater burden of severe influenza-associated illness, hospitalization, and mortality than younger populations [1]. During the 2018–19 and 2019–20 influenza seasons in the United States (US), the numbers of total estimated influenza-related hospitalizations and deaths were highest in the ≥ 65 age group [2,3]. Older adults with underlying medical conditions are particularly vulnerable to serious influenza illness and complications [4,5].

Vaccination is the most effective means of preventing influenza infection and associated sequelae, and the Advisory Committee on Immunization Practices (ACIP) in the US considers older adults a priority vaccination group [6]. A Cochrane review of randomized clinical trials found that influenza vaccination in older adults may reduce influenza infections by 58% compared to unvaccinated populations [7]. However, vaccination in older adults is less effective at preventing severe outcomes due to frailty, increased presence of comorbidities, and immunosenescence [5,8–10]. Indeed, a meta-analysis of test‐negative design case–control studies between 2010 and 2015 found that the pooled vaccine effectiveness against any type of influenza was 51% for adults 18–64 years of age and only 37% for older adults [11].

Adding a vaccine adjuvant is one approach to enhancing the immune response of a vaccinated individual. During the 2018–19 and 2019–20 influenza seasons in the US, available enhanced vaccines approved for adults ≥65 years included adjuvanted trivalent inactivated influenza vaccine (Fluad®, Seqirus; aIIV3) [12,13]. aIIV3 combines the MF59 adjuvant (a squalene-based oil-in-water emulsion) and a standard dose of antigen to provide a stronger, more durable immune response than standard-dose vaccines [14,15]. Real-world evidence (RWE) evaluating clinical outcomes among those ≥65 years has shown a superior clinical benefit of aIIV3 over standard egg-derived quadrivalent inactivated influenza vaccines (IIV4e) [16–21]. In a meta-analysis of RWE from cohort design studies, the pooled estimate for the relative vaccine effectiveness (rVE) of aIIV3 compared with IIV4e for the prevention of influenza-related medical encounters was 13.7% (95% confidence interval [CI]: 3.1%–24.2%) [21]. As a result of the evidence supporting the use of enhanced influenza vaccines in an older adult population, the ACIP recommended that adults ≥65 years preferentially receive enhanced higher dose or adjuvanted influenza vaccines during the 2022–23 influenza season [6].

Despite the clinical evidence supporting vaccination with aIIV3 compared to IIV4e, economic outcomes associated with the two vaccines have not been compared in an older adult population in the US leveraging the use of real-world data. We aimed to investigate whether the ACIP clinical recommendation favoring enhanced vaccines, and specifically aIIV3, can be supported from an economic perspective. The current study quantified the potential cost savings associated with the receipt of aIIV3 compared to IIV4e for the 2018–19 and 2019–20 influenza seasons. The clinical benefit of aIIV3 versus IIV4e based on real-world data was anticipated to translate to a lower economic burden due to prevented hospitalizations related to influenza and associated complications. This study followed a similar methodology as a previously published economic assessment of enhanced vaccines [22].

2. Patients and methods

2.1. Study design and data sources

This study was conducted in two stages. First, we conducted a retrospective cohort study using deidentified data from several IQVIA databases: Professional Fee Claims (Dx), Longitudinal Prescription Claims (LRx), and Hospital Charge Data Master (CDM). Clinical outcomes were compared between aIIV3 and IIV4e using inverse probability of treatment weighting (IPTW) to adjust for imbalances in measured confounders and treatment selection bias [23]. An economic assessment was then undertaken to quantify potential cost savings due to avoided cardiorespiratory and respiratory hospitalizations through vaccination with aIIV3 instead of IIV4e using the real-world data from the retrospective cohort study. Hospitalizations and associated costs were not directly compared in the retrospective analysis due to the potential for confounding factors, following a previously published approach [22].

The study databases have been previously described and utilized to compare clinical and economic outcomes associated with different influenza vaccines among older adults [20,24,25]. Dx includes professional claims representing approximately 70% of physician activity in the US. LRx includes more than 1.6 B retail and mail-order prescriptions, representing dispensed prescriptions for approximately 85% of all pharmacies in the US. CDM includes records for over 400 hospitals, covering 7 M annual inpatient stays and 60 M annual outpatient visits [25]. The datasets were linked using a deterministic matching algorithm using actual patient information, helping to ensure continuity of patient follow-up across databases [26]. The data were in compliance with the Health Insurance Portability and Accountability Act (HIPAA). Ethical review and approval were not required for the retrospective analysis of the deidentified secondary data.

2.2. Study population

Adults aged ≥65 years with a claim for aIIV3 or IIV4e in the LRx or Dx databases were identified via National Drug Codes (NDCs) and Current Procedural Terminology (CPT) codes during the vaccination window and assigned to two mutually exclusive cohorts. Sample populations were defined separately for the 2018–19 and 2019–20 influenza seasons. The 2018–19 influenza season was defined from 1 August 2018 to 31 July 2019, with a vaccination window from 1 August 2018 to 31 January 2019 and a study period from 1 February 2018 to 31 July 2019 [18,24]. The 2019–20 influenza season was defined from 4 August 2019 to 7 March 2020, with a vaccination window from 4 August 2019 to 31 January 2020 and a study period from 4 February 2019 to 7 March 2020 [19,25]. The shorter influenza season for 2019–20 was chosen to avoid outcome misclassification associated with the coronavirus disease 2019 (COVID-19) pandemic. Note that IIV4e contains two influenza B antigens (one of each B lineage), while aIIV3 includes only one [12,13]. However, there was very little influenza B activity during the 2018–19 influenza season, and allV3 matched the predominant circulating lineage (B/Victoria) during the 2019–20 influenza season [12,27,28].

The first claim for aIIV3 or IIV4e during the respective vaccination window was termed the index date and determined the vaccine cohort. Patients were required to have linkage in both the Dx and LRx databases at any time during the respective study period. Because these are open-source databases, proxies for continuous enrollment (CE) were applied to establish the 6-month preindex (baseline) period and variable postindex (follow-up) period through the end of the influenza season. These requirements included ≥ 1 claim in both databases in the 6 months prior to the 6-month preindex period and in the 6 months following the end of the respective study period and pharmacy stability, defined as consistent reporting of data from the pharmacy most frequently visited by the patient in each month from the start of the 6-month preindex period to the end of the study period (as reported monthly for each pharmacy in the LRx database).

The exclusion criteria included ≥ 1 influenza-related hospitalization, emergency room visit, or outpatient office visit from the start of the vaccination window up to 13 days after the index date, based on a diagnosis code for influenza (International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM] 487.x, 488.x; International Classification of Diseases, Tenth Revision, Clinical Modification [ICD-10-CM] J09.x, J10.x, J11.x [29]) in any position, >1 influenza vaccine during the influenza season, missing demographic data (e.g. sex, region, or payer type), or lack of linkage to CDM. For the 2019–20 influenza season, those with ≥ 1 claim with a COVID-19 diagnosis (ICD-10-CM U07.1) during the study period were additionally excluded to avoid the potential impact of the COVID-19 pandemic on the clinical outcomes of respiratory-related and cardiorespiratory-related hospitalizations and associated costs.

2.3. Study measures

2.3.1. Patient characteristics

Patient demographics were assessed on the index date, while baseline clinical characteristics, including the Charlson Comorbidity Index (CCI; Dartmouth-Manitoba adaptation [30,31]), comorbid conditions, and indicators of health-seeking behavior, were measured during the 6-month preindex period (not including the index date) and identified through ICD-9-CM and ICD-10-CM diagnosis codes in the Dx and CDM databases. Indicators of frail health status were also assessed via ICD-9-CM and ICD-10-CM diagnosis codes and procedure codes and included home oxygen use, wheelchair use, walker use, dementia, urinary catheter use, falls, and fractures. Baseline healthcare resource utilization (HCRU) and costs were measured over the 6-month preindex period. As Dx and CDM capture charges and not costs, a cost-to-charge ratio (CCR) was applied [32,33]. Costs were reported in 2021 US dollars (USD) for this study.

2.3.2. Adjusted clinical outcomes

Clinical outcomes included respiratory-related and cardiorespiratory-related hospitalizations, which were identified from both Dx and CDM and measured over the variable postindex period, starting from the index date +14 days to the end of the respective influenza seasons. The 14-day period after the index date was included to allow for an immunological response following vaccination. Hospitalizations with an admission date before the end of the respective influenza seasons but with a discharge date after were considered for the outcome assessment. Cardiorespiratory-related hospitalizations were defined as hospitalizations with a primary discharge diagnosis code of ICD-9-CM 390.x-519.x or ICD-10-CM Ixx-Jxx. Respiratory-related hospitalizations were defined as hospitalizations with a primary discharge diagnosis code of ICD-9-CM 460.x-519.x or ICD-10-CM Jxx. As influenza vaccination is not expected to prevent falls or fractures, fall- and fracture-related hospitalizations, defined as hospitalizations with a primary discharge diagnosis for a fall or fracture event, were also evaluated as a negative control outcome. For each clinical outcome of interest, the first occurring event per subject was included. An individual could contribute an event for more than one outcome (e.g. cardiorespiratory-related hospitalization and respiratory-related hospitalization) but could not contribute more than one event to the same outcome.

2.3.3. Economic outcomes

For the economic assessment, the total number of respiratory-related and cardiorespiratory-related hospitalizations, as well as the cost per respiratory-related or cardiorespiratory-related hospitalization, were assessed. Hospitalization costs (applying the CCR to charges) were taken from CDM only, as CDM more closely captures hospitalization costs.

2.4. Statistical analyses

All analyses were conducted separately for each influenza season. To account for potential measured confounding between the aIIV3 and IIV4e cohorts, propensity scores (PSs) were calculated using logistic regression with the vaccine cohort as the dependent variable and transformed into weights for IPTW adjustment. This approach has been previously described by the authors in further detail [20,24,25]. IPTW was used to estimate the average treatment effect (ATE; i.e. expected causal effect of the treatment across all individuals in the population) and to create a weighted sample (or ‘pseudopopulation’) with a balanced distribution of measured covariates between cohorts. A stabilized IPTW approach was utilized, which reduced the type I error [23,34]. Additionally, weight values greater than 10 were truncated to 10 given the potential bias of outliers.

Baseline characteristics with a standardized mean difference (SMD) ≥0.10 between cohorts prior to adjustment were considered imbalanced and included in the IPTW logistic regression models [35]. For both the 2018–19 and 2019–20 samples, this included age, payer type, index month, U. Department of Health and Human Services (DHHS) region, and CCI category. Diabetes was also included in the 2018–19 sample. The SMD was calculated to evaluate the performance of the weighting procedure, and an SMD < 0.10 between cohorts was considered adequately balanced [35].

For the clinical outcomes of interest, IPTW-weighted multivariate Poisson regression models were used to estimate adjusted rate ratios (RRs) and 95% CIs. In addition to IPTW, the models adjusted for clinically relevant variables that were not included in the IPTW logistic regression models: sex, any frailty indicator (yes/no) and any preindex hospitalization (yes/no). The adjusted rVE for aIIV3 vs. IIV4e was calculated as ([1-RR] * 100%). All analyses were based on observed data without projection. SAS® Release 9.4 (SAS Institute Inc., Cary, NC, USA) was utilized for the analyses.

2.5. Economic assessment

The economic assessment followed the methodology from a previous publication evaluating higher dose (HD)-IIV3 and aIIV3 [22]. Cost savings associated with respiratory-related and cardiorespiratory-related hospitalizations were evaluated separately for each influenza season. The total number of respiratory and cardiorespiratory hospitalizations, as observed from both cohorts combined, were assigned to aIIV3 and IIV4e recipients based on the respective adjusted rVE, and the incidence rate was calculated. The absolute risk reduction (ARR) was then derived by subtracting the incidence rate in the aIIV3 cohort from the incidence rate in the IIV4e cohort. The inverse of the ARR was taken, which resulted in the number needed to vaccinate (NNV = 1/ARR). The NNV is the estimated number of patients who need to receive aIIV3 vaccination instead of IIV4e to prevent one additional hospitalization event.

Cost savings associated with aIIV3 vaccination (i.e. net cost savings per aIIV3 recipient) were estimated by calculating the difference in costs per IIV4e recipient as if they had been vaccinated with aIIV3 instead. The average cost of observed hospitalizations (i.e. the cost of a prevented hospitalization with aIIV3) was divided by the NNV, and then the average difference in cost of the two vaccines was subtracted. The wholesale acquisition cost (WAC) difference between aIIV3 and IIV4e was $30.69 in 2018–19 and $34.86 in 2019–20 [36]. Finally, we calculated the total realized net cost savings as the total number of aIIV3 recipients multiplied by the cost savings per aIIV3 recipient.

2.6. Sensitivity analysis

Clinical and economic outcomes were measured over the respective high-influenza activity period (HIAP) for each influenza season as a sensitivity analysis. The observation period was restricted from later of ([index date + 14] or start of the HIAP), to the end of the HIAP. For the 2018–19 influenza season, the HIAP was defined from 23 December 2018 to 30 March 2019, and for the 2019–20 influenza season, it was defined from 8 December 2019 to 7 March 2020. These HIAPs were determined through a moving epidemic method (MEM) algorithm to establish epidemic thresholds for the start and end of the influenza season [37,38].

3. Results

3.1. Sample and baseline characteristics

For the 2018–19 influenza season, a total of 715,807 aIIV3 and 320,991 IIV4e recipients were included in this study (Figure 1). Prior to adjustment, a few variables were imbalanced between cohorts (Table S1). aIIV3 patients were older (75.2 vs. 74.0 years), and more patients used Medicare (49% vs. 38%) or Medicare Part D (30% vs. 17%) compared to IIV4e patients. There were also regional differences, and more aIIV3 patients were in the South (48.6% vs. 40.1%). Fewer aIIV3 patients were vaccinated in November (11.6% vs. 15.6%). The mean CCI score (0.9 vs. 1.1) and proportion with diabetes (21% vs. 26%) were lower for the aIIV3 cohort. After IPTW adjustment, all baseline demographic clinical characteristics, HCRU, and costs were well balanced (Table 1).

Figure 1. Patient selection.

For the 2018-19 influenza season, the vaccination window was defined from 1 August 2018 to 31 January 2019, and the study period was from 1 February 2018 to 31 July 2019. For the 2019-20 influenza season, the vaccination window was defined from 4 August 2019 to 31 January 2020, and the study period was defined from 4 February 2019 to 7 March 2020.

aIIV3 = adjuvanted trivalent influenza vaccine; CDM = Hospital Charge Data Master; COVID-19 = coronavirus disease 2019; Dx = Professional Fee Claims; ER = emergency room; IIV4e = standard egg-derived quadrivalent influenza vaccine; LRx = Longitudinal Prescription Claims.

Table 1. Demographic and preindex (6 months) clinical characteristics, HCRU and costs, post-IPTW, 2018–2019.

	Adjusted^a
	aIIV3 (N = 716,279)	IIV4e (N = 320,197)	SMD^b
Demographic Characteristics
Age (years), mean (SD)	74.8 (6.6)	74.9 (7.1)	−0.01
Female, n (%)	420,170 (58.7%)	193,845 (60.5%)	0.04
Payer type, n (%)
Cash	3,292 (0.5%)	1,468 (0.5%)	0.00
Medicaid	763 (0.1%)	308 (0.1%)	0.00
Medicare Part D	187,009 (26.1%)	83,323 (26%)	0.00
Medicare	325,005 (45.4%)	145,474 (45.4%)	0.00
Third party	200,211 (28%)	89,624 (28%)	0.00
Geographic region, n (%)
Northeast	130,584 (18.2%)	58,734 (18.3%)	0.00
Midwest	115,364 (16.1%)	52,348 (16.4%)	0.01
South	332,387 (46.4%)	144,998 (45.3%)	−0.02
West	137,945 (19.3%)	64,117 (20%)	0.02
DHHS region, n (%)
Region 1: CT, ME, MA, NH, RI, VT	40,482 (5.7%)	18,510 (5.8%)	0.01
Region 2: NJ, NY, PR, VI	68,338 (9.5%)	30,689 (9.6%)	0.00
Region 3: DE, DC, MD, PA, VA, WV	67,174 (9.4%)	31,080 (9.7%)	0.01
Region 4: AL, FL, GA, KY, MS, NC, SC, TN	209,970 (29.3%)	91,787 (28.7%)	−0.01
Region 5: IL, IN, MI, MN, OH, WI	99,494 (13.9%)	44,785 (14%)	0.00
Region 6: AR, LA, NM, OK, TX	83,011 (11.6%)	37,418 (11.7%)	0.00
Region 7: IA, KS, MO, NE	15,464 (2.2%)	6,946 (2.2%)	0.00
Region 8: CO, MT, ND, SD, UT, WY	7,541 (1.1%)	3,412 (1.1%)	0.00
Region 9: AZ, CA, HI, NV, AS, FS, GU, PU	89,720 (12.5%)	39,792 (12.4%)	0.00
Region 10: AK, ID, OR, WA	35,085 (4.9%)	15,777 (4.9%)	0.00
Clinical Characteristics
Month of influenza vaccination, n (%)
August	26,035 (3.6%)	11,666 (3.6%)	0.00
September	196,982 (27.5%)	85,501 (26.7%)	−0.02
October	351,392 (49.1%)	158,637 (49.5%)	0.01
November	92,859 (13%)	42,073 (13.1%)	0.01
December	30,926 (4.3%)	14,111 (4.4%)	0.00
January	18,085 (2.5%)	8,209 (2.6%)	0.00
CCI score, n (%)
0	384,880 (53.7%)	171,036 (53.4%)	−0.01
1	151,467 (21.2%)	68,012 (21.2%)	0.00
2	90,907 (12.7%)	40,892 (12.8%)	0.00
3	39,982 (5.6%)	18,044 (5.6%)	0.00
4+	49,045 (6.9%)	22,213 (6.9%)	0.00
CCI score, mean (SD)	1 (1.4)	1 (1.4)	0.01
Preindex comorbid conditions, n (%)
Asthma	28,177 (3.9%)	13,316 (4.2%)	0.01
Blood disorders	2,195 (0.3%)	1,062 (0.3%)	0.00
Chronic lung disease	62,702 (8.8%)	30,947 (9.7%)	0.03
Diabetes	163,496 (22.8%)	73,562 (23%)	0.00
Heart disease	94,540 (13.2%)	44,254 (13.8%)	0.02
Kidney disorders	67,924 (9.5%)	28,756 (9%)	−0.02
Liver disorders	18,051 (2.5%)	7,898 (2.5%)	0.00
Neurological or neurodevelopmental conditions	36,137 (5.1%)	17,662 (5.5%)	0.02
Weakened immune system^c	73,674 (10.3%)	32,333 (10.1%)	−0.01
IBD	4,803 (0.7%)	2,127 (0.7%)	0.00
Composite of the above	353,156 (49.3%)	159,869 (49.9%)	0.01
Frail health status, n (%)
Home oxygen use	33,450 (4.7%)	17,178 (5.4%)	0.03
Wheelchair use	18,560 (2.6%)	10,104 (3.2%)	0.03
Walker use	25,378 (3.5%)	12,263 (3.8%)	0.02
Dementia	9,649 (1.4%)	4,156 (1.3%)	0.00
Urinary catheter use	2,004 (0.3%)	767 (0.2%)	−0.01
Falls	5,524 (0.8%)	2,176 (0.7%)	−0.01
Fractures	3,400 (0.5%)	1,359 (0.4%)	−0.01
Composite of the above	78,103 (10.9%)	38,283 (12%)	0.03
Indicators of health-seeking behavior (n, %)
Cataracts	64,634 (9%)	26,872 (8.4%)	−0.02
Eyelid disorders	9,586 (1.3%)	4,026 (1.3%)	−0.01
Hemorrhoids	15,011 (2.1%)	6,513 (2%)	0.00
Ingrown nail	7,922 (1.1%)	3,702 (1.2%)	0.00
Lipomas	1,832 (0.3%)	797 (0.3%)	0.00
UTI	45,538 (6.4%)	22,643 (7.1%)	0.03
Wound of hand or finger	3,209 (0.5%)	1,463 (0.5%)	0.00
Composite of the above	134,736 (18.8%)	60,157 (18.8%)	0.00
Preindex HCRU and Costs
Any all-cause hospitalization, n (%)	57,061 (8.0%)	25,873 (8.1%)	0.00
Total all-cause costs, mean (SD)	$6,906 ($20,739)	$6,803 ($21,152)	−0.01
All-cause outpatient pharmacy costs, mean (SD)	$2,025 ($6,272)	$2,152 ($7,068)	0.04
All-cause medical costs^d
Inpatient, mean (SD)	$900 ($8,381)	$808 ($8,289)	0.00
ER, mean (SD)	$174 ($988)	$169 ($976)	0.02
Outpatient medical, mean (SD)	$3,808 ($15,675)	$3,674 ($15,282)	−0.03
Outpatient physician office visits, mean (SD)	$1,419 ($7,783)	$1,266 ($7,372)	0.00

^aIPTW was used to adjust for imbalances in measured confounding between the aIIV3 and IIV4e cohorts. The following variables were used in the IPTW adjustment: age (mean), payer type, index month, DHHS region, CCI category and diabetes.

^bAn SMD (absolute) ≥ 0.10 was considered imbalanced.

^cIncluding HIV/AIDS; metastatic cancer and acute leukemia; lung or upper digestive or other severe cancer; lymphatic, head, neck, brain, or major cancer; breast, prostate, colorectal, or other cancer; and immune system disorders.

^dMedical costs were calculated using the national cost to charge ratio.

aIIV3 = adjuvanted trivalent influenza vaccine; CCI = Charlson Comorbidity Index; DHHS = US Department of Health & Human Services; HCRU = healthcare resource utilization; IBD = inflammatory bowel disease; IIV4e = standard egg-derived quadrivalent influenza vaccine; IPTW = inverse probability of treatment weighting; SD = standard deviation; SMD = standardized mean difference; UTI = urinary tract infection.

In the 2019–20 influenza season, a total of 844,169 aIIV3 and 306,270 IIV4e recipients were included (Figure 1). Similar to the 2018–19 influenza season, prior to adjustment, aIIV3 patients were older (75.2 years vs. 74.1 years), included more patients with Medicare (41% vs. 35%) or Medicare Part D (35% vs. 21%) and were more often in the South (51.5% vs. 41.1%) compared to IIV4e patients (Table S1). Fewer aIIV3 patients were vaccinated in October (41.1% vs. 53.9%). The mean CCI score was lower for the aIIV3 cohort (1.0 vs. 1.1). After adjustment, all baseline demographic and clinical characteristics, HCRU, and costs were well balanced (Table 2).

Table 2. Demographic and preindex (6 months) clinical characteristics, HCRU and costs, post-IPTW, 2019–2020.

	Adjusted^a
	aIIV3 (N = 845,069)	IIV4e (N = 304,050)	SMD^b
Demographic Characteristics
Age (years), mean (SD)	74.9 (6.4)	74.9 (6.9)	−0.02
Female, n (%)	494,765 (58.6%)	183,429 (60.3%)	0.04
Payer type, n (%)
Cash	2,497 (0.3%)	914 (0.3%)	0.00
Medicaid	893 (0.1%)	333 (0.1%)	0.00
Medicare Part D	264,142 (31.3%)	93,535 (30.8%)	−0.01
Medicare	334,107 (39.5%)	119,250 (39.2%)	−0.01
Third party	243,430 (28.8%)	90,018 (29.6%)	0.02
Geographic region, n (%)
Northeast	137,343 (16.3%)	48,612 (16%)	−0.01
Midwest	147,320 (17.4%)	54,074 (17.8%)	0.01
South	413,961 (49%)	147,506 (48.5%)	−0.01
West	146,445 (17.3%)	53,858 (17.7%)	0.01
DHHS region, n (%)
Region 1: CT, ME, MA, NH, RI, VT	40,387 (4.8%)	14,286 (4.7%)	0.00
Region 2: NJ, NY, PR, VI	71,436 (8.5%)	25,668 (8.4%)	0.00
Region 3: DE, DC, MD, PA, VA, WV	78,777 (9.3%)	28,777 (9.5%)	0.00
Region 4: AL, FL, GA, KY, MS, NC, SC, TN	264,992 (31.4%)	95,339 (31.4%)	0.00
Region 5: IL, IN, MI, MN, OH, WI	128,169 (15.2%)	46,774 (15.4%)	0.01
Region 6: AR, LA, NM, OK, TX	101,569 (12%)	36,652 (12.1%)	0.00
Region 7: IA, KS, MO, NE	18,687 (2.2%)	6,749 (2.2%)	0.00
Region 8: CO, MT, ND, SD, UT, WY	8,769 (1%)	3,243 (1.1%)	0.00
Region 9: AZ, CA, HI, NV, AS, FS, GU, PU	95,047 (11.3%)	33,495 (11%)	−0.01
Region 10: AK, ID, OR, WA	37,236 (4.4%)	13,068 (4.3%)	−0.01
Clinical Characteristics
Month of influenza vaccination, n (%)
August	71,949 (8.5%)	25,796 (8.5%)	0.00
September	237,623 (28.1%)	83,484 (27.5%)	−0.01
October	376,897 (44.6%)	136,996 (45.1%)	0.01
November	104,619 (12.4%)	38,141 (12.5%)	0.01
December	36,217 (4.3%)	13,224 (4.4%)	0.00
January	17,763 (2.1%)	6,409 (2.1%)	0.00
CCI score, n (%)
0	437,195 (51.7%)	156,263 (51.4%)	−0.01
1	180,295 (21.3%)	65,021 (21.4%)	0.00
2	111,832 (13.2%)	40,566 (13.3%)	0.00
3	50,638 (6%)	18,370 (6%)	0.00
4+	65,110 (7.7%)	23,831 (7.8%)	0.01
CCI score, mean (SD)	1 (1.4)	1 (1.5)	0.01
Preindex comorbid conditions, n (%)
Asthma	34,494 (4.1%)	13,068 (4.3%)	0.01
Blood disorders	2,800 (0.3%)	1,011 (0.3%)	0.00
Chronic lung disease	79,615 (9.4%)	30,881 (10.2%)	0.02
Diabetes	196,354 (23.2%)	73,303 (24.1%)	0.02
Heart disease	120,975 (14.3%)	45,042 (14.8%)	0.01
Kidney disorders	89,957 (10.6%)	30,483 (10%)	−0.02
Liver disorders	23,122 (2.7%)	8,186 (2.7%)	0.00
Neurological or neurodevelopmental conditions	45,100 (5.3%)	17,380 (5.7%)	0.02
Weakened immune system^c	95,103 (11.3%)	31,331 (10.3%)	−0.03
IBD	5,926 (0.7%)	2,012 (0.7%)	0.00
Composite of the above	436,164 (51.6%)	157,480 (51.8%)	0.00
Frail health status, n (%)
Home oxygen use	44,930 (5.3%)	18,480 (6.1%)	0.03
Wheelchair use	24,194 (2.9%)	10,677 (3.5%)	0.04
Walker use	32,444 (3.8%)	12,462 (4.1%)	0.01
Dementia	12,635 (1.5%)	4,412 (1.5%)	0.00
Urinary catheter use	1,798 (0.2%)	597 (0.2%)	0.00
Falls	7,182 (0.9%)	2,371 (0.8%)	−0.01
Fractures	4,533 (0.5%)	1,545 (0.5%)	0.00
Composite of the above	101,877 (12.1%)	39,882 (13.1%)	0.03
Indicators of health-seeking behavior (n, %)
Cataracts	71,798 (8.5%)	23,891 (7.9%)	−0.02
Eyelid disorders	10,983 (1.3%)	3,483 (1.2%)	−0.01
Hemorrhoids	17,458 (2.1%)	6,036 (2%)	−0.01
Ingrown nail	8,605 (1%)	3,150 (1%)	0.00
Lipomas	2,192 (0.3%)	719 (0.2%)	0.00
UTI	55,522 (6.6%)	22,157 (7.3%)	0.03
Wound of hand or finger	3,845 (0.5%)	1,472 (0.5%)	0.00
Composite of the above	155,964 (18.5%)	55,653 (18.3%)	0.00
Preindex HCRU and Costs
Any all-cause hospitalization, n (%)	71,034 (8.5%)	25,647 (8.4%)	0.00
Total all-cause costs, mean (SD)	$7,612 ($21,687)	$7,228 ($21,587)	−0.03
All-cause outpatient pharmacy costs, mean (SD)	$2,259 ($6,769)	$2,338 ($7,894)	0.02
All-cause medical costs^d
Inpatient, mean (SD)	$988 ($9,507)	$868 ($8,804)	0.00
ER, mean (SD)	$211 ($1,195)	$199 ($1,129)	0.01
Outpatient medical, mean (SD)	$4,154 ($15,694)	$3,824 ($15,403)	−0.05
Outpatient physician office visits, mean (SD)	$1,580 ($8,991)	$1,320 ($7,511)	−0.02

^aIPTW was used to adjust for imbalances in measured confounding between the aIIV3 and IIV4e cohorts. The following variables were used in the IPTW adjustment: age (mean), payer type, index month, DHHS region, and CCI score (continuous).

^bAn SMD (absolute) ≥ 0.10 was considered imbalanced.

^cIncluding HIV/AIDS; metastatic cancer and acute leukemia; lung or upper digestive or other severe cancer; lymphatic, head, neck, brain, or major cancer; breast, prostate, colorectal, or other cancer; and immune system disorders.

^dMedical costs were calculated using the national cost to charge ratio.

aIIV3 = adjuvanted trivalent influenza vaccine; CCI = Charlson Comorbidity Index; DHHS = US Department of Health & Human Services; HCRU = healthcare resource utilization; IBD = inflammatory bowel disease; IIV4e = standard egg-derived quadrivalent influenza vaccine; IPTW = inverse probability of treatment weighting;SD = standard deviation; SMD = standardized mean difference; UTI = urinary tract infection.

3.2. Adjusted clinical outcomes

For both influenza seasons, after IPTW adjustment and multivariable Poisson regression, aIIV3 was significantly more effective than IIV4e in preventing hospitalizations related to cardiorespiratory disease (2018–19 rVE = 6.2%; 95% CI: 4.7%–7.7%; and 2019–20 rVE = 6.0%; 95% CI: 4.0%–7.9%) (Figure 2; Table S2). aIIV3 was also significantly more effective than IIV4e in preventing hospitalizations related to respiratory disease (2018–19 rVE = 8.9%; 95% CI: 6.5%–11.2%; and 2019–20 rVE = 10.1%; 95% CI: 7.1%–13.0%). The rVE for fall/fracture-related hospitalizations for both influenza seasons was not significant, further suggesting that the cohorts were well-balanced with IPTW.

Figure 2. Adjusted (post-IPTW and Poisson regression) rVE of aIIV3 vs. IIV4e at preventing hospitalization.

aIIV3 = adjuvanted trivalent influenza vaccine; CI = confidence interval; IIV4e = standard egg-derived quadrivalent influenza vaccine; IPTW = inverse probability of treatment weighting; rVE = relative vaccine effectiveness.

3.3. Economic assessment

3.3.1. Cardiorespiratory hospitalization cost savings

During the 2018–19 influenza season, the total number of observed cardiorespiratory hospitalizations in the study sample was 270,002, with mean costs per hospitalization of $34,282 (Table 3). During the 2019–20 influenza season, there were 159,943 total observed cardiorespiratory hospitalizations, with a mean cost per hospitalization of $34,135.

Table 3. Total prevented hospitalizations and cost savings associated with aIIV3 vs. IIV4e.

	Influenza Season 2018–19	Influenza Season 2019–20
N	1,036,798	1,150,439
aIIV3	715,807	844,169
IIV4e	320,991	306,270
Cardiorespiratory hospitalizations
Observed hospitalizations, n	270,002	159,943
Cost per hospitalization, mean (SD)	$34,282 ($50,896)	$34,135 ($46,090)
rVE (95% CI)	6.21% (4.67% − 7.72%)	5.97% (3.96% − 7.94%)
aIIV3 hospitalization rate^a (per 10,000 patients) (95% CI)	2,552 (2,538–2,565)	1,367 (1,359–1,375)
IIV4e hospitalization rate^a (per 10,000 patients) (95% CI)	2,721 (2,691–2,751)	1,454 (1,432–1,476)
ARR^b (95% CI)	168.9 (136–201.6)	86.8 (63.1–110.3)
NNV^c (95% CI)	59.2 (49.6–73.6)	115.2 (90.7–158.5)
Net cost savings per IIV4e recipient^d (95% CI)	$548 ($435–$660)	$261 ($180–$342)
Total realized cost savings for this population^e (95% CI)	$392,431,074 ($311,642,724–$472,756,571)	$220,667,884 ($152,341,345–$288,357,112)
Respiratory hospitalizations
Observed hospitalizations, n	104,971	63,206
Cost per hospitalization, mean (SD)	$24,347 ($50,427)	$24,896 ($39,266)
rVE (95% CI)	8.87% (6.49% − 11.19%)	10.11% (7.07% − 13.04%)
aIIV3 hospitalization rate^a (per 10,000 patients) (95% CI)	983 (974–991)	533 (528–538)
IIV4e hospitalization rate^a (per 10,000 patients) (95% CI)	1,078 (1,060–1,097)	593 (579–608)
ARR^b (95% CI)	95.6 (75.1–116)	60 (44.9–74.8)
NNV^c (95% CI)	104.6 (86.2–133.1)	166.7 (133.7–222.5)
Net cost savings per IIV4e recipient^d (95% CI)	$202 ($152–$252)	$114 ($77–$151)
Total realized cost savings for this population^e (95% CI)	$144,696,237 ($108,942,412–$180,137,597)	$96,628,579 ($65,028,756–$127,785,809)

^aCalculated by applying the adjusted rVE to the total observed number of hospitalizations across both vaccination cohorts.

^bThe ARR was calculated by subtracting the hospitalization rate in the aIIV3 cohort from the hospitalization rate in the IIV4e cohort.

^cThe NNV was calculated by taking the multiplicative inverse of the AAR (NNV = 1/AAR), i.e. the number of patients who needed to be vaccinated with aIIV3 instead of IIV4e to prevent one additional hospitalization.

^dRefers to the difference in costs per IIV4e recipient if they received aIIV3 instead. Calculated as the average cost of all observed hospitalizations for IIV4e recipients divided by the NNV, minus the average cost difference of the two vaccines.

^eCalculated by multiplying the total number of aIIV3 recipients by the net cost savings per aIIV3 recipient.

aIIV3 = adjuvanted trivalent influenza vaccine; ARR = absolute risk reduction; CI = confidence interval; IIV4e = standard egg-derived quadrivalent influenza vaccine; N = number of patients; NNV = number needed to vaccinate; rVE = relative vaccine effectiveness.

Applying the adjusted rVE, the cardiorespiratory hospitalization rates for aIIV3 and IIV4e recipients were 2,552 (95% CI: 2,538–2,565) and 2,721 (95% CI: 2,691–2,751) per 10,000 person-years (PY), respectively, for the 2018–19 influenza season and 1,367 (95% CI: 1,359–1,375) and 1,454 (95% CI: 1,432–1,476) per 10,000 PY, respectively, for the 2019–20 influenza season.

During the 2018–19 influenza season, the NNV with allV3 was 59.2 (95% CI: 49.6–73.6) (Figures 3 and 4; Table 3). Net cost savings per aIIV3 recipient due to avoided cardiorespiratory hospitalizations was $548 (95% CI: $435–$660). The total realized cost savings for this population was $392 M (95% CI: $312 M–$472 M). During the 2019–20 influenza season, the NNV with aIIV3 was 115.2 (95% CI: 90.7–158.5), resulting in net cost savings per aIIV3 recipient of $261 (95% CI: $180–$342) due to avoided cardiorespiratory hospitalizations and a total realized cost savings for this population of $221 M (95% CI: $152 M–$288 M).

Figure 3. Net cost savings per aIIV3 recipient in USD, estimated by calculating the difference in costs per IIV4e recipient as if they had received aIIV3 instead.

aIIV3 = adjuvanted trivalent influenza vaccine; CI = confidence interval; IIV4e = standard egg-derived quadrivalent influenza vaccine.

Figure 4. Total realized cost savings in millions (USD) from prevented cardiorespiratory and respiratory hospitalizations with aIIV3 vaccination instead of IIV4e vaccination.

aIIV3 = adjuvanted trivalent influenza vaccine; CI = confidence interval; IIV4e = standard egg-derived quadrivalent influenza vaccine.

3.3.2. Respiratory hospitalization cost savings

During the 2018–19 influenza season, the total number of observed respiratory hospitalizations in the study sample was 104,971, with mean costs per hospitalization of $24,347 (Table 3). During the 2019–20 influenza season, there were 63,206 total observed respiratory hospitalizations, with mean costs per hospitalization of $24,896. Applying the adjusted rVE, the respiratory hospitalization rates for aIIV3 and IIV4e recipients were 983 (95% CI: 974–991) and 1,078 (95% CI: 1,060–1,097) per 10,000 PY, respectively, for the 2018–19 influenza season and 533 (95% CI: 528–538) and 593 (95% CI: 579–608) per 10,000 PY, respectively, for the 2019–20 influenza season.

During the 2018–19 influenza season, the NNV with aIIV3 was 104.6 (95% CI: 86.2–133.1) (Table 3). Net cost savings per aIIV3 recipient due to avoided respiratory hospitalizations was $202 (95% CI: $152–$252). The total realized cost savings for this population was $145 M (95% CI: $109 M–$180 M). During the 2019–20 influenza season, the NNV with aIIV3 was 166.7 (95% CI: 133.7–222.5), resulting in net cost savings per aIIV3 recipient of $114 (95% CI: $77–$151) due to avoided respiratory hospitalizations and a total realized cost savings for this population of $97 M (95% CI: $65 M–$128 M).

3.3.3. Sensitivity analysis

When limiting analyses to the HIAP only, aIIV3 continued to be significantly effective in preventing both cardiorespiratory and respiratory hospitalizations compared to IIV4e during both the 2018–19 and 2019–20 HIAPs (Table S3). However, because there were fewer hospitalizations during this shorter assessment period and less potential for cost savings with aIIV3, estimated cost savings per aIIV3 recipient and in total were lower compared to the main analysis. Net cost savings for avoided cardiorespiratory hospitalizations during the 2018–19 and 2019–20 HIAPs were $185 (95% CI: $123–$247) and $239 (95% CI: $172–$304) per aIIV3 recipient, with total realized cost savings of $132 M (95% CI: $88 M–$177 M) and $201 M (95% CI: $145 M–$257 M), respectively (Table S4). Net cost savings for avoided respiratory hospitalizations during the 2018–19 and 2019–20 HIAP seasons were $65 (95% CI: $38–$92) and $96 (95% CI: $63–$128) per aIIV3 recipient, with total realized cost savings of $47 M (95% CI: $27 M–$66 M) and $81 M (95% CI: $53 M–$108 M), respectively.

4. Discussion

This is one of the first studies to compare economic outcomes between older adult patients vaccinated with aIIV3 or IIV4e leveraging real-world data. For both the 2018–19 and 2019–20 influenza seasons, aIIV3 was significantly more effective than IIV4e in preventing cardiorespiratory and respiratory hospitalizations compared to IIV4e. These real-world clinical benefits were quantified and translated to substantial cost savings for aIIV3 recipients. In our study population, aIIV3 was associated with an estimated $392 M and $145 M in total realized cost savings compared to IIV4e in terms of avoided cardiorespiratory and respiratory hospitalizations, respectively, in the 2018–19 influenza season and $221 M and $97 M in the 2019–20 influenza season, respectively. It is important to consider that these estimated cost savings were specific to the observed study populations. Additionally, only cost savings related to prevented cardiorespiratory and respiratory hospitalizations were captured in our study; therefore, we likely underestimated the true cost savings. Our 2019–20 influenza season assessment was also abbreviated due to the COVID-19 pandemic.

Our real-world clinical findings are supported by prior studies showing the superior rVE of aIIV3 compared to IIV4 [16–19,21]. In a prior US study linking electronic medical records (EMRs) and claims data, researchers assessed the rVE of aIIV3 with that of IIV4 in preventing influenza-related medical encounters (inpatient and outpatient). They found that compared to IIV3, aIIV3 was significantly more effective at preventing influenza-related medical encounters during the 2018–19 influenza season (rVE = 27.8%; 95% CI: 25.7%–29.9%) [16]. The rVE was greater than in our study, which might be in part related to the different outcome definitions and data sources. In the present study, our primary outcome of interest was cardiorespiratory and respiratory hospitalizations as opposed to any influenza-related medical encounter. Furthermore, the requirement of linkage to CDM in our study may have biased the study sample toward a more severely ill population due to the requirement of having activity in a hospital. Indeed, in a separate study by the same authors leveraging the same data sources but which required the population to have at least one health condition, researchers found the rVE of aIIV3 versus IIV4 to be lower at 20.4% (95% CI: 16.2%–24.4%) for the 2018–19 influenza season [17]. Other studies in Medicare fee-for-service populations have similarly shown a superior rVE of aIIV3 compared to IIV4 at preventing influenza-related hospitalizations in older adult populations in both the 2018–19 and 2019–20 influenza seasons [18,19].

Economic comparisons of aIIV3 compared to IIV4 are limited, and this was highlighted by two recent literature reviews on the cost-effectiveness of influenza vaccines in older adults [39,40]. Few studies were found with a pairwise comparison reported between aIIV3 and IIV4. In these studies, developed models (two in Italy and one in the US) found aIIV3 to be consistently dominant compared to IIV4 due to lower cost and greater health benefits [40].

In any assessment of influenza vaccines, it is important to consider that the epidemiology of influenza varies each year in terms of timing, duration, activity, and clinical severity. During the 2018–19 influenza season, which was described as having moderate severity, influenza A viruses predominated, with very little influenza B activity. Influenza A(H1N1)pdm09 viruses dominated from October 2018 through mid-February 2019, and beginning in late February 2019, influenza A(H3N2) clade 3C.3a viruses were predominant. This antigenic drift resulted in a mismatch between the circulating virus strains and vaccine virus strains [28]. In the 2019–20 influenza season, also described as having moderate severity, influenza activity in the US began to increase in November and was consistently high throughout January and February. The season was characterized by two waves of activity beginning with influenza B viruses followed by A(H1N1)pdm09 viruses [3]. As noted earlier, B/Victoria viruses predominated, matching the lineage included in aIIV3 that year [12,27]. Influenza activity declined in March 2020, which may have been the result of community prevention measures for COVID-19 [40]. Importantly, our findings showed a superior rVE and economic benefits of aIIV3 compared to IIV4 over two epidemiologically different influenza seasons. This is consistent with the literature, as aIIV3 has been shown to produce a stronger and more durable immune response over a broader range of virus strains than conventional standard dose vaccines [15,41,42].

Given the seasonal variability of influenza epidemiology, these results provide important clinical and economic data. The findings presented in this study complement the recent ACIP recommendations by demonstrating that in addition to the clinical rationale, there is an economic justification for vaccination with adjuvanted vs. standard egg-derived influenza vaccines despite the added acquisition cost. We provide new economic evidence supporting this clinical recommendation by showing both the clinical and economic advantages associated with aIIV3. While our findings are specific to the US, we believe they have global implications. For other countries where enhanced vaccines are not recommended over standard dose vaccines, potentially related to economic factors, we demonstrate that enhanced vaccines can be cost-effective due to the prevention of hospitalizations. This is also relevant from a public health point of view. Our findings may help inform influenza vaccine recommendations in other countries. Importantly, for the 2020–21 influenza season, a quadrivalent formulation of the adjuvanted influenza vaccine was introduced (aIIV4), and aIIV3 is no longer available in the US [43]. aIIV3 provided limited immunity against circulating B strains of the lineage not contained in the vaccine [44]. With the introduction of aIIV4, the adjuvanted vaccine now contains both lineages of influenza B [43].

These findings should be interpreted with consideration of several limitations. First, although IPTW was used to adjust for imbalances in measured confounding, it is possible that residual confounding may have remained due to unmeasured confounding. Nevertheless, our sample was well balanced across baseline characteristics post-IPTW, including indicators of frailty and health-seeking behavior, and no significant difference was observed in the negative control outcome of falls/fractures. Second, there are limitations related to the use of open-source claims databases. Data are utilized as available from a patient but are not considered comprehensive. Data are obtained from pharmacies, offices and hospitals that participate in and contribute to the database, and there is no insight into activity at pharmacies, offices, and hospitals that do not contribute to the data. Nevertheless, LRx and Dx coverage in the US is relatively high. We applied proxies for continuous enrollment based on patient activity and pharmacy reporting to provide greater assurance that patient healthcare activity within these databases was captured during the study period. Third, mandating linkage to CDM at any point in time during the availability of CDM data may have biased the study sample toward a more severe population due to the requirement of having activity in a hospital (inpatient or outpatient setting) in CDM. These database-related limitations were generally expected to apply to all identified patients from both vaccine cohorts equally, without differential impact.

Additionally, there may be limitations in the economic assessment due to limitations associated with cost data. First, only charges are available in Dx and CDM (prior to adjudication) and not true paid costs; thus, a national cost to charge ratio was applied to charges. Second, our economic assessment was limited to the effect of both vaccines on cardiorespiratory and respiratory-related hospitalizations and associated costs; however, the reduction of vaccine-preventable influenza is likely to provide cost savings beyond the hospital setting, both direct and indirect, including at other healthcare settings and reduced workplace productivity loss. Therefore, total cost savings may be higher than we estimated in this study. Third, the savings in our study related to differences in vaccine cost were based on vaccine WAC and may not necessarily reflect market prices or prices for specific populations. Finally, the cost savings identified in this assessment were specific to the observed study populations and are not generalizable to other populations or to the national US level. However, our study also has considerable strengths. Our economic assessment utilized real-world data, which we adjusted for confounders using robust methodology. As discussed above, our cost savings findings associated with aIIV4 were consistent across two epidemiologically different influenza seasons, as well as over HIAPs. In addition, our data sources are representative of the 65+ population in the US and include all payers.

5. Conclusions

Over both the 2018–19 and 2019–20 influenza seasons, aIIV3 was significantly more effective than IIV4e in preventing hospitalizations related to cardiorespiratory and respiratory disease. Reduced hospitalizations with receipt of aIIV3 compared to IIV4e translated into substantial cost savings in our study populations. Our findings suggest that aIIV3 provides not only superior clinical benefits but also economic benefits versus IIV4e in an older adult population. We anticipate observing even greater clinical and economic benefits in future influenza seasons with aIIV4 compared to IIV4e.

Declaration of interest

V Divino, Z Zhao and M DeKoven are employees of IQVIA, which received funding for this study from Seqirus. SI Pelton, MJ Postma and M Levin received financial support for their time and effort from Seqirus for this study. J Mould-Quevedo is an employee of CSL Seqirus Inc. and a shareholder. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or material discussed in the manuscript apart from those disclosed.

Reviewer disclosures

A reviewer on this manuscript has received grant funding from Sanofi and has received funding from Merck. They serve on an advisory board for GSK. Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Author contributions

Conceptualization, JM, VD, MD; study design and data interpretation, JM, VD, MD, ZZ, SIP, MJP and MJL; drafting of the article, VD; critical revision of the article, SIP, JM, MD, ZZ, MJP and MJL; funding acquisition, JM. All authors have read and provided final approval of the submitted version of the manuscript. All authors attest that they meet the ICMJE criteria for authorship.

Ethics approval

Ethical review and approval were not required for the retrospective analysis of the deidentified secondary data. Informed consent was not required for this type of study.

Geolocation information

United States

Acknowledgments

The authors would like to thank Lia Pizzicato, Senior Consultant at IQVIA, for medical writing assistance in the preparation of this paper.

Data availability statement

The original deidentified data used in this analysis were obtained from and are the property of IQVIA. IQVIA has restrictions prohibiting the authors from making the dataset publicly available. Interested researchers may contact IQVIA to apply to gain access to this study’s data in the same way the authors obtained the data (see https://www.iqvia.com/contact/sf).

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/14760584.2023.2293237.

Additional information

Funding

This study was funded by CSL Seqirus Inc.