- Email: [email protected]
Public health impact of Rotarix vaccination among commercially insured children in the United States
Vaccine xxx (2017) xxx–xxx
Contents lists available at ScienceDirect
Vaccine journal homepage: www.elsevier.com/locate/vaccine
Public health impact of Rotarix vaccination among commercially insured children in the United States Girishanthy Krishnarajah a,1, Andrew Kageleiry b, Caroline Korves b, Patrick Lefebvre c, Mei S. Duh b,⇑ a
GSK, 5 Crescent Drive, Philadephia, PA 19112 United States Analysis Group, Inc., 111 Huntington Avenue, 14th Floor, Boston, MA 02199, United States c Groupe d’analyse, Ltée, 1000 De La Gauchetière Ouest, Bureau 1200, Montréal, Quebec H3B 4W5, Canada b
a r t i c l e
i n f o
Article history: Received 16 February 2017 Received in revised form 25 May 2017 Accepted 10 June 2017 Available online xxxx Keywords: Rotavirus Rotarix Diarrhea Gastroenteritis Commercially insured
a b s t r a c t Background: This study (NCT01915888) assessed public health impact of Rotarix, GSK [RV1] vaccination. Methods: Children born between 2007–2011 were identified from Truven Commercial Claims and Encounters Databases and observed until earlier of plan disenrollment or five years old. Children receiving one or two doses of RV1 during the vaccination window were assigned to incomplete and complete vaccination cohorts, respectively. Children without rotavirus (RV) vaccination (RV1 OR RotaTeq, Merck & Co., Inc. [RV5]) were assigned to the unvaccinated cohort. Claims with International Classification of Disease 9th edition (ICD-9) codes for diarrhea and RV infections were identified. First RV episode incidence, RV-related and diarrhea-related healthcare resource utilization were compared. Multivariate Poisson regression with generalized estimating equations was used to generate 95% confidence intervals (CIs) around incidence rate ratios (IRR) between cohorts while adjusting for gender, age and calendar year. Mean costs for first RV and diarrhea episodes were calculated with adjustment for gender and birth year; bootstrapping was used to determine statistically significant differences between cohorts. Results: Incidence of first RV episodes was significantly reduced in complete and incomplete vaccination cohorts compared to the unvaccinated cohort (IRR = 0.17 [95%CI: 0.09–0.30] and IRR = 0.19 [95%CI: 0.06– 0.58], respectively). RV-related inpatient, outpatient and emergency room (ER) visits were significantly lower for complete vaccination versus unvaccinated cohort. Diarrhea-related inpatient and ER visit rates were significantly lower for complete vaccination versus unvaccinated cohorts; outpatient rates were similar. RV-related and diarrhea-related resource utilization rates were significantly lower or no different for incomplete vaccination versus unvaccinated cohort. Compared with unvaccinated children, adjusted mean cost for first RV episode and first diarrhea episode per 1000 persons was $11,511 (95%CI: $9855$12,024) and $46,772 (95%CI: $26,268-$66,604) lower, respectively, for completely vaccinated children. Conclusions: RV1 vaccination confers benefits in reduction of RV incidence, RV- and diarrhea-related healthcare resource utilization, and RV- and diarrhea-related healthcare costs. Ó 2017 GlaxoSmithKline. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
1. Introduction Before initiation of a routine infant rotavirus (RV) vaccination program in February 2006 in the United States (US), nearly every Abbreviations: RV, rotavirus; RV1, Rotarix, GSK; RV5, RotaTeq, Merck & Co., Inc.; ICD-9, International Classification of Disease 9th edition; CI, confidence interval; ACIP, Advisory Committee on Immunization Practices; ER, emergency room; IR, incidence rate; IRR, incidence rate ratio; PI, package insert. ⇑ Corresponding author. E-mail addresses: [email protected] (G. Krishnarajah), Andrew. [email protected], [email protected] (A. Kageleiry), Caroline. [email protected] (C. Korves), [email protected] (P. Lefebvre), [email protected] (M.S. Duh). 1 Present address: CSL Behring, 1020 1st Avenue, King of Prussia, Pennsylvania 19406 United States.
child was infected with RV by age 5 years [1]. Since then, both the Advisory Committee on Immunization Practices (ACIP) and the American Academy of Pediatrics have recommended routine immunization of children with RV vaccine [1,2]. Two brands of oral RV vaccines are approved for children in the US by its Food and Drug Administration: RotaTeq, Merck & Co., Inc. (RV5) (approved in 2006) and Rotarix, GSK (RV1) (approved in 2008) [1]. RV1 is administered in two doses; the vaccine is administered in three doses if RV5 or a mix of brands (i.e., both RV1 and RV5) is used. The ACIP recommends a vaccination window from 6 weeks old to 8 months old [1]. The package insert (PI) for RV1 specifies the first of two doses should be administered to children beginning at 6 weeks of age, and the second should be administered with at least a four week gap from the first dose and before 24 weeks of age [3].
http://dx.doi.org/10.1016/j.vaccine.2017.06.034 0264-410X/Ó 2017 GlaxoSmithKline. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
Please cite this article in press as: Krishnarajah G et al. Public health impact of Rotarix vaccination among commercially insured children in the United States. Vaccine (2017), http://dx.doi.org/10.1016/j.vaccine.2017.06.034
2
G. Krishnarajah et al. / Vaccine xxx (2017) xxx–xxx
Payers need to evaluate the impact of RV vaccination on both clinical and economic endpoints from a real-world perspective. RV1 and RV5, with their different antigen compositions and dosing schedules [4], are both currently in use in the US. Published studies to date using real-world practice data in the US to evaluate effectiveness of complete and incomplete vaccination have focused only on RV5 [5–7] or both RV1 and RV5 [8–10]. Studies on vaccine effectiveness in real world settings provide important public health data and can contribute to decision-making about ongoing vaccination programs. The present study attempts to address current gaps of knowledge by estimating the impact of RV1 vaccination on RV incidence and healthcare resource utilization among children aged less than 5 years using commercial insurance claims data. 2. Methods 2.1. Data source Data from the 2007–2011 Truven Commercial Claims and Encounters database were analyzed. Data from the 2007–2011 Truven Medicaid database were analyzed, but the sample size was inadequate for valid statistical comparisons as the cohorts for complete vaccination and incomplete vaccination were approximately 2000, and there was only one RV episode per each cohort; these results are not presented. Truven databases are described in detail in Krishnarajah et al. [11]. The current study (NCT01915888) utilized 2007–2011 data captured from approximately 9 million children under 5 years of age with commercial insurance coverage. 2.2. Patient selection Children were further required to have continuously received medical and pharmacy benefits since birth until end of follow-up (defined as disenrollment from insurance plans, 5 years of age, death, or data cutoff, whichever occurred first) and not to have enrolled in capitation-based health plans. With capitation-based health plans, participating physicians are paid a pre-determined amount of money per patient enrolled for a specified range of services [12]. As there is no fee for these services, administrative databases may have incomplete records for utilization and cost. Children residing in states with universal vaccination programs were excluded from the analysis as done in prior studies [10]; vaccinations in these states were not likely billed to insurance companies; thus, no record of vaccination would appear in the database. Children receiving any RV5 vaccination during follow-up were excluded. 2.3. Study cohorts Children were assigned to cohorts based on RV1 vaccines received. Individuals’ person-time was assessed from 6 weeks through 24 weeks old to determine RV1 vaccination exposure. RV1 vaccination was identified by current procedural terminology code 90681 or national drug code 58160-805-01, 58160-805-02, 58160-805-11, 58160-851-01, 58160-853-02, and 58160-854-52. As insurance claims data do not include date of birth, the authors provided Truven Health Analytics, Inc. with dates of service for RV1 vaccination. Truven Health Analytics, Inc., with access to exact date of birth data, identified whether vaccination was given during the 6–24 week age window. Children given two RV1 vaccinations during the vaccination window at least four weeks apart contributed person-time to the complete vaccination cohort; those with one RV1 vaccination contributed to the incomplete vaccination cohort; children without vaccination contributed to the
unvaccinated cohort. The unvaccinated cohort was comprised of individuals who reached 24 weeks of age old on or after January 1, 2008; originally, January 1, 2006 was the cut-off date, but this was changed as only RV5 (and not RV1) was available in the 2006–2008 timeframe, and non-vaccination during the time of RV1 availability was of interest. 2.4. Outcomes Incidence rate ratios (IRR) of the first RV episode were estimated to compare rates between cohorts. The first occurrence of an RV-coded encounter marked incidence of first RV episode. RVcoded encounters were identified by claims with the specific International Classification of Diseases 9th Edition (ICD-9) code for RV (008.61 - Enteritis due to rotavirus). A single RV episode included groups of RV-related medical claims dated less than 28 days apart. Children who had the first RV episode prior to 24 weeks of age did not contribute to the RV incidence estimation. The rate of RV-related healthcare resource utilization was calculated with the numerator being the number of RV-related hospitalizations, outpatient visits or emergency room (ER) visits (including visits to ER or urgent care facility) occurring on different calendar days, and the denominator being the total follow-up time since end of vaccination window within a cohort. Costs per first RV episode were calculated as the sum of all medical claims dated within the 14 days before the first RV-coded claim and the 14 days after the last RV-coded claim within an RV episode where any RVcoded claims occurring within a 14 day period were considered to be the same episode [9,11]. An individual without an RV episode thus had a cost of $0. Because laboratory tests confirming RV infection are not performed routinely on all patients with diarrhea, claims coded as RV likely represent only a fraction of all RV-related encounters [5,13]; therefore, diarrhea incidence was also estimated. Diarrhea-coded healthcare encounters were identified by claims with non-specific ICD-9 codes for diarrhea. See Supplementary Table 1 for details. Similar methods were used to calculate rate of diarrhea-related visits and costs for first diarrhea episodes. 2.5. Statistical analysis Analyses were conducted according to PI recommendations, and frequencies and proportions were reported to describe baseline characteristics of the study cohorts. Pearson’s chi-square test was used to compare differences between cohorts. Poisson regression models with generalized estimating equations were used to estimate IRRs and 95% confidence intervals (CIs) for RV- and diarrhea-related outcomes. Both univariate Poisson regressions and post hoc multivariate Poisson regressions were performed. Since the likelihood of RV infection varies with age and calendar year [14], in part due to the change in prevalence of the virus which is affected by vaccination, multivariate analyses were necessary to account for potential confounding by these factors as they were significantly different between cohorts. Mean cost per 1000 children per first RV episode was calculated for each cohort by summing episode costs for children within a cohort and dividing that by the number of children within the cohort, and then multiplying that by 1000. The difference in mean cost per 1000 children per first RV episode was then calculated by subtracting the mean of one cohort from that of another [9,11]. Univariate and post hoc multivariate linear regression (with adjustment for gender and birth year) were performed. Cost data are often not normally distributed [15]. Thus, p-values and 95% CIs for cost differences between vaccination cohorts were computed using bootstrapping (with replacement) with 999 iterations. Similarly, diarrhearelated resource use and costs of the first diarrhea episode were
Please cite this article in press as: Krishnarajah G et al. Public health impact of Rotarix vaccination among commercially insured children in the United States. Vaccine (2017), http://dx.doi.org/10.1016/j.vaccine.2017.06.034
3
G. Krishnarajah et al. / Vaccine xxx (2017) xxx–xxx
estimated and compared between cohorts. Costs were inflationadjusted to 2012 US dollars based on the medical care component of the Consumer Price Index. All analyses were performed using SAS software version 9.3 (SAS Institute, Inc., Cary, NC, USA).
3. Results A total of 225,587 children born between 2007 and 2011 with commercial insurance were eligible for the analysis (Table 1). Of
Table 1 Characteristics of eligible children. Complete vaccination [A]
Incomplete vaccination [B]
Unvaccinated [C]
p-values A vs. B
A vs. C
B vs. C
Overall study period (2008–2011)
N = 34,928
N = 8390
N = 182,269
Year of birth, n (%) 2008 2009 2010 2011
3392 (9.7%) 17,574 (50.3%) 13,571 (38.9%) 391 (1.1%)
913 (10.9%) 4146 (49.4%) 3246 (38.7%) 85 (1.0%)
65,465 (35.9%) 49,655 (27.2%) 39,934 (21.9%) 655 (0.4%)
0.001 0.139 0.780 0.401
<0.001 <0.001 <0.001 <0.001
<0.001 <0.001 <0.001 <0.001
Gender, n (%) Female
16,871 (48.3%)
4054 (48.3%)
88,137 (48.4%)
0.977
0.855
0.949
Type of health plan, n (%) Comprehensive Preferred provider organization Other
187 (0.5%) 26,514 (75.9%) 8227 (23.6%)
49 (0.6%) 6314 (75.3%) 2027 (24.2%)
1969 (1.1%) 140,502 (77.1%) 39,798 (21.8%)
0.587 0.209 0.241
<0.001 <0.001 <0.001
<0.001 <0.001 <0.001
Region, n (%) Northeast North Central South West Unknown
4819 (13.8%) 10,310 (29.5%) 14,902 (42.7%) 4320 (12.4%) 577 (1.7%)
1199 (14.3%) 2118 (25.2%) 3793 (45.2%) 1162 (13.8%) 118 (1.4%)
22,110 (12.1%) 57,016 (31.3%) 72,823 (40.0%) 27,018 (14.8%) 3302 (1.8%)
0.240 <0.001 <0.001 <0.001 0.108
<0.001 <0.001 <0.001 <0.001 0.039
<0.001 <0.001 <0.001 0.014 0.006
9173 (26.3%)
2213 (26.4%)
40,471 (22.2%)
0.831
<0.001
<0.001
4493 3139 3589 3778
1208 (14.4%) 865 (10.3%) 948 (11.3%) 1025 (12.2%)
23,079 15,317 16,459 18,352
<0.001 <0.001 0.006 <0.001
0.300 <0.001 <0.001 <0.001
<0.001 <0.001 <0.001 <0.001
3542 (10.1%) 2110 (6.0%)
853 (10.2%) 540 (6.4%)
11,722 (6.4%) 10,250 (5.6%)
0.944 0.175
<0.001 0.002
<0.001 0.002
1908 (5.5%) 985 (2.8%)
531 (6.3%) 287 (3.4%)
9758 (5.4%) 5990 (3.3%)
0.002 0.003
0.408 <0.001
<0.001 0.500
1799 (5.2%)
447 (5.3%)
7493 (4.1%)
0.511
<0.001
<0.001
Year 2008 Age at start of year, n (%) Born in year 0–1 years 1–2 years 2–3 years 3–4 years
N = 181
N = 123
N = 61,719
181 (100.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%)
123 (100.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%)
35,159 (57.0%) 26,560 (43.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%)
1.000 1.000 1.000 1.000 1.000
<0.001 <0.001 1.000 1.000 1.000
<0.001 <0.001 1.000 1.000 1.000
Year 2009 Age at start of year, n (%) Born in year 0–1 years 1–2 years 2–3 years 3–4 years
N = 14,373
N = 3132
N = 107,979
11,006 (76.6%) 3367 (23.4%) 0 (0.0%) 0 (0.0%) 0 (0.0%)
2238 (71.5%) 894 (28.5%) 0 (0.0%) 0 (0.0%) 0 (0.0%)
29,049 (26.9%) 59,521 (55.1%) 19,409 (18.0%) 0 (0.0%) 0 (0.0%)
<0.001 <0.001 1.000 1.000 1.000
<0.001 <0.001 <0.001 1.000 1.000
<0.001 <0.001 <0.001 1.000 1.000
Year 2010 Age at start of year, n (%) Born in year 0–1 years 1–2 years 2–3 years 3–4 years
N = 22,494
N = 5577
N = 107,971
6256 (27.8%) 14,086 (62.6%) 2152 (9.6%) 0 (0.0%) 0 (0.0%)
1632 (29.3%) 3396 (60.9%) 549 (9.8%) 0 (0.0%) 0 (0.0%)
22,451 (20.8%) 39,014 (36.1%) 34,763 (32.2%) 11,743 (10.9%) 0 (0.0%)
0.031 0.017 0.530 1.000 1.000
<0.001 <0.001 <0.001 <0.001 1.000
<0.001 <0.001 <0.001 <0.001 1.000
Year 2011 Age at start of year, n (%) Born in year 0–1 years 1–2 years 2–3 years 3–4 years
N = 23,987
N = 5668
N = 92,417
391 (1.6%) 11,973 (49.9%) 10,076 (42.0%) 1547 (6.4%) 0 (0.0%)
85 (1.5%) 2834 (50.0%) 2369 (41.8%) 380 (6.7%) 0 (0.0%)
655 (0.7%) 33,340 (36.1%) 26,236 (28.4%) 23,873 (25.8%) 8313 (9.0%)
0.482 0.908 0.773 0.484 1.000
<0.001 <0.001 <0.001 <0.001 <0.001
<0.001 <0.001 <0.001 <0.001 <0.001
Comorbidities, n (%) Acute upper respiratory infections of multiple or unspecified sites Suppurative and unspecified otitis media Acute bronchitis and bronchiolitis General symptoms Symptoms involving respiratory system and other chest symptoms Diseases of esophagus Viral and chlamydial infection in conditions classified elsewhere and of unspecified site Symptoms involving digestive system Nonsuppurative otitis media and Eustachian tube disorders Atopic dermatitis and related conditions
(12.9%) (9.0%) (10.3%) (10.8%)
(12.7%) (8.4%) (9.0%) (10.1%)
Please cite this article in press as: Krishnarajah G et al. Public health impact of Rotarix vaccination among commercially insured children in the United States. Vaccine (2017), http://dx.doi.org/10.1016/j.vaccine.2017.06.034
4
G. Krishnarajah et al. / Vaccine xxx (2017) xxx–xxx
those who reached 24 weeks of age on or after January 1, 2008, 182,269 (80.8%) served as the unvaccinated cohort; 34,928 (15.5%) were classified as completely vaccinated; the remaining 8390 (3.7%) were grouped as the incomplete vaccination cohort. Year of birth varied significantly across cohorts (Table 1). In the unvaccinated cohort, 35.9% of individuals were born in 2008 whereas approximately 10% in the vaccination cohorts were born that year (p-value <0.001 for complete vaccination versus unvaccinated, and for incomplete vaccination versus unvaccinated). The proportion of patients in the vaccinated cohorts born during 2009 and 2010 was nearly double that of the unvaccinated cohort. Thus, for this study, the vaccinated cohorts had more observation time among younger children and in later calendar years relative to the unvaccinated cohort. Following the vaccination window, there were 11, 3 and 449 RV episodes observed in the complete vaccination, incomplete vaccination and unvaccinated cohorts, respectively (Supplementary Table 2). Using the multivariate model adjusting for gender, age and calendar year, the IRR for RV episodes for the complete vaccination cohort and incomplete vaccination cohorts versus the unvaccinated cohort were 0.17 (95% CI: 0.09–0.30) and 0.19 (95% CI: 0.06–0.58), respectively (Table 2). In the multivariate analysis of RV-related resource utilization (Supplementary Table 3 and Table 3), complete vaccination versus no vaccination was associated with significant reductions in inpatient visits (IRR = 0.07 [95% CI: 0.02–0.21]), outpatient visits (IRR = 0.25 [95% CI: 0.11–0.54]), and ER visits (IRR = 0.09 [95% CI: 0.03–0.32]). Compared with the unvaccinated cohort, children
with incomplete vaccination also had lower rates of RV-related inpatient (IRR = 0.09 [95% CI: 0.01–0.67]), and ER visits (IRR = 0.19 [95% CI: 0.05–0.78]); there was no significant difference for outpatient visits (IRR = 0.41 [95% CI: 0.09–1.94]). When comparing rates of diarrhea-related medical encounters between children with complete vaccination versus no vaccination, those completely vaccinated had statistically significantly lower rates of inpatient visits (IRR = 0.37 [95% CI: 0.29–0.47]), and ER visits (IRR = 0.62 [95% CI: 0.57–0.68]); there was no significant difference in outpatient visits (IRR = 1.02 [95% CI: 0.98–1.06]) (Supplementary Table 4 and Table 4). Compared with the unvaccinated cohort, children with incomplete vaccination had similar rates of diarrhea-related inpatient (IRR = 0.78 [95% CI: 0.52– 1.16]) and outpatient visits (IRR = 1.01 [95% CI: 0.94–1.08]). Incompletely vaccinated children had a statistically significantly lower rate of ER visits relative to the unvaccinated (IRR = 0.83 [95% CI: 0.71–0.98]) (Supplementary Table 4 and Table 4). Adjusted mean total healthcare cost per 1000 persons for first RV episode in the complete vaccination cohort was significantly lower than that in the unvaccinated cohort (difference = $11,511 [95% CI: $12,024 to $9855]) (Table 5). Similar cost difference was observed between the incomplete vaccination and unvaccinated cohorts (difference = $9809 [95% CI: $11,511 to $5413]). Across comparisons, differences were largely driven by statistically significant differences in inpatient visit costs. The mean costs of individual first diarrhea episodes per 1000 persons were much higher than the mean costs of the first RV episodes per 1000 persons, mainly because of the much higher incidence
Table 2 Comparison of incidence of first RV episode between vaccinated and unvaccinated cohorts. p-value
Adjusted IRR (95% CI)a
p-value
<0.001
0.17 (0.09, 0.30) Reference
<0.001
Gender Female Male
0.80 (0.67, 0.96) Reference
0.019
Age (years) 0–1 1–2 2–3 3–4
1.45 (1.18, 1.77) 1.52 (1.25, 1.85) 0.73 (0.56, 0.95) Reference
<0.001 <0.001 0.022
Year 2008 2009 2010 2011
0.99 (0.80, 1.23) 1.25 (1.08, 1.44) 0.39 (0.32, 0.48) Reference
0.920 0.003 <0.001
0.19 (0.06, 0.58) Reference
0.004
Gender Female Male
0.81 (0.67, 0.97) Reference
0.023
Age (years) 0–1 1–2 2–3 3–4
1.47 (1.20, 1.80) 1.55 (1.28, 1.89) 0.71 (0.54, 0.93) Reference
<0.001 <0.001 0.012
Year 2008 2009 2010 2011
0.98 (0.79, 1.22) 1.23 (1.06, 1.42) 0.40 (0.32, 0.49) Reference
0.878 0.005 <0.001
Unadjusted IRR (95% CI) Complete vaccination versus unvaccinated RV1 vaccination cohort Complete vaccination 0.18 (0.10, 0.33) Unvaccinated Reference
Incomplete vaccination versus unvaccinated RV1 vaccination cohort Incomplete vaccination 0.21 (0.07, 0.64) Unvaccinated Reference
0.006
Abbreviations: IRR, incidence rate ratio; CI, confidence interval; RV, rotavirus; RV1, Rotarix, GSK. a Adjusted for age, calendar year and gender.
Please cite this article in press as: Krishnarajah G et al. Public health impact of Rotarix vaccination among commercially insured children in the United States. Vaccine (2017), http://dx.doi.org/10.1016/j.vaccine.2017.06.034
5
G. Krishnarajah et al. / Vaccine xxx (2017) xxx–xxx Table 3 Comparison of RV-related resource utilization between vaccinated and unvaccinated cohorts. IRR (95% CI)
p-value
IRR (95% CI)
p-value
Complete vaccination versus unvaccinated Inpatient visits Outpatient visits Unadjusted RV1 vaccination cohort Complete vaccination Unvaccinated Adjusteda RV1 vaccination cohort Complete vaccination Unvaccinated Gender Female Male Age (years) 0–1 1–2 2–3 3–4 Year 2008 2009 2010 2011
Adjusteda RV1 vaccination cohort Incomplete vaccination Unvaccinated Gender Female Male Age (years) 0–1 1–2 2–3 3–4 Year 2008 2009 2010 2011
p-value
ER visits
0.08 (0.03, 0.26) Reference
<0.001
0.25 (0.13, 0.47) Reference
<0.001
0.11 (0.04, 0.31) Reference
<0.001
0.07 (0.02, 0.21) Reference
<0.001
0.25 (0.11, 0.54) Reference
<0.001
0.09 (0.03, 0.32) Reference
<0.001
0.90 (0.71, 1.15) Reference
0.412
0.66 (0.49, 0.89) Reference
0.007
0.86 (0.65, 1.15) Reference
0.316
1.59 (1.20, 2.11) 1.93 (1.47, 2.53) 0.69 (0.48, 1.01) Reference
0.001 <0.001 0.054
1.19 (0.87, 1.63) 1.12 (0.84, 1.50) 0.74 (0.50, 1.10) Reference
0.266 0.445 0.138
1.34 (1.00, 1.79) 1.50 (1.16, 1.94) 0.64 (0.44, 0.94) Reference
0.051 0.002 0.023
0.87 (0.65, 1.17) 1.10 (0.91, 1.33) 0.43 (0.33, 0.55) Reference
0.359 0.335 <0.001
1.02 (0.74, 1.41) 1.28 (1.02, 1.60) 0.43 (0.32, 0.59) Reference
0.901 0.036 <0.001
0.74 (0.49, 1.11) 1.21 (0.95, 1.54) 0.41 (0.30, 0.57) Reference
0.143 0.121 <0.001
Incomplete vaccination versus unvaccinated Inpatient visits Outpatient visits Unadjusted RV1 vaccination cohort Incomplete vaccination Unvaccinated
IRR (95% CI)
ER visits
0.12 (0.02, 0.82) Reference
0.031
0.42 (0.16, 1.13) Reference
0.086
0.24 (0.06, 0.97) Reference
0.045
0.09 (0.01, 0.67) Reference
0.018
0.41 (0.09, 1.94) Reference
0.260
0.19 (0.05, 0.78) Reference
0.021
0.92 (0.72, 1.17) Reference
0.480
0.64 (0.47, 0.88) Reference
0.005
0.88 (0.66, 1.17) Reference
0.365
1.58 (1.19, 2.10) 1.95 (1.48, 2.55) 0.70 (0.48, 1.01) Reference
0.002 <0.001 0.056
1.22 (0.89, 1.67) 1.16 (0.86, 1.55) 0.71 (0.48, 1.06) Reference
0.210 0.337 0.095
1.34 (1.00, 1.80) 1.55 (1.19, 2.00) 0.62 (0.42, 0.92) Reference
0.048 <0.001 0.016
0.87 (0.65, 1.17) 1.09 (0.90, 1.32) 0.43 (0.33, 0.55) Reference
0.373 0.387 <0.001
1.01 (0.73, 1.39) 1.25 (0.99, 1.57) 0.44 (0.32, 0.60) Reference
0.952 0.059 <0.001
0.74 (0.50, 1.11) 1.18 (0.93, 1.50) 0.41 (0.30, 0.57) Reference
0.143 0.173 <0.001
Abbreviations: IRR, incidence rate ratio; CI, confidence interval; RV1, Rotarix, GSK; ER, emergency room. a Adjusted for age, calendar year and gender.
of diarrhea-related episodes relative to RV episodes. Across inpatient, outpatient, and ER visit settings, mean costs for first RV episode were uniformly statistically significantly lower for the complete and incomplete vaccination cohorts compared with the unvaccinated cohort. 4. Discussion Findings from the current study suggest that vaccination with RV1 results in significant reduction in RV-coded visits for children following the vaccination window. Higher RV-related healthcare resource utilization and costs were also observed in the unvaccinated cohort versus children with complete RV1 vaccination. Findings from this current study are in agreement with other published studies on RV vaccination with RV5 or mixed vaccination [5–10]. Also compatible with previous findings, this study found lower RV-coded encounters among individuals with incomplete RV vaccination versus those in the unvaccinated cohort [9,16]. A prior study of children under 5 years of age with commercial and Medicaid insurance showed incomplete vaccination with either RV1 or
RV5 versus no vaccination was associated with statistically significant reductions in first RV episodes and RV-related healthcare utilization [9]. A study of RV5 exclusively showed incomplete vaccination versus no vaccination was associated with reductions in both RV-related and diarrhea-related hospitalizations and ER visits [16]. Notably, all of these studies are observational and have relied on insurance claims data. The current study was not designed to evaluate incomplete RV1 vaccination, and incomplete RV1 vaccination has not been evaluated in clinical trials and is not recommended by ACIP. In the current analysis, the complete and incomplete vaccination cohorts were significantly younger than the unvaccinated cohort; approximately 40% of children included in the vaccinated cohorts were born in 2010 and 2011 versus only 22% in the unvaccinated cohort. This observation is consistent with reports that uptake of RV vaccination in general improved in the first several years following availability; nationally, RV vaccination coverage among 19–35 month olds in the US rose from 43.9% in 2009 to 67.3% in 2011 [17] and was reported to be 73.2% in 2015 [18]. In the current study, children aged 2 years and
Please cite this article in press as: Krishnarajah G et al. Public health impact of Rotarix vaccination among commercially insured children in the United States. Vaccine (2017), http://dx.doi.org/10.1016/j.vaccine.2017.06.034
6
G. Krishnarajah et al. / Vaccine xxx (2017) xxx–xxx
Table 4 Comparison of diarrhea-related resource utilization between vaccinated and unvaccinated cohorts. IRR (95% CI)
p-value
IRR (95% CI)
p-value
Complete vaccination versus unvaccinated Inpatient visits Outpatient visits Unadjusted RV1 vaccination cohort Complete vaccination Unvaccinated Adjusteda RV1 vaccination cohort Complete vaccination Unvaccinated Gender Female Male Age (years) 0–1 1–2 2–3 3–4 Year 2008 2009 2010 2011
Adjusteda RV1 vaccination cohort Incomplete vaccination Unvaccinated Gender Female Male Age (years) 0–1 1–2 2–3 3–4 Year 2008 2009 2010 2011
p-value
ER visits
0.43 (0.34, 0.54) Reference
<0.001
1.09 (1.06, 1.12) Reference
<0.001
0.70 (0.65, 0.76) Reference
<0.001
0.37 (0.29, 0.47) Reference
<0.001
1.02 (0.98, 1.06) Reference
0.299
0.62 (0.57, 0.68) Reference
<0.001
0.82 (0.73, 0.92) Reference
0.001
0.86 (0.84, 0.88) Reference
<0.001
0.84 (0.80, 0.89) Reference
<0.001
1.49 (1.32, 1.69) 1.43 (1.27, 1.61) 0.67 (0.57, 0.79) Reference
<0.001 <0.001 <0.001
1.41 (1.37, 1.45) 1.33 (1.30, 1.37) 0.82 (0.80, 0.85) Reference
<0.001 <0.001 <0.001
1.45 (1.37, 1.53) 1.29 (1.22, 1.36) 0.79 (0.74, 0.85) Reference
<0.001 <0.001 <0.001
0.88 (0.78, 1.00) 0.98 (0.90, 1.07) 0.81 (0.73, 0.88) Reference
0.058 0.633 <0.001
1.01 (0.98, 1.04) 1.00 (0.99, 1.02) 0.96 (0.94, 0.97) Reference
0.440 0.626 <0.001
0.81 (0.76, 0.87) 1.04 (1.00, 1.09) 0.89 (0.86, 0.93) Reference
<0.001 0.081 <0.001
Incomplete vaccination versus unvaccinated Inpatient visits Outpatient visits Unadjusted RV1 vaccination cohort Incomplete vaccination Unvaccinated
IRR (95% CI)
ER visits
0.88 (0.63, 1.21) Reference
0.420
1.08 (1.03, 1.14) Reference
0.002
0.95 (0.83, 1.09) Reference
0.448
0.78 (0.52, 1.16) Reference
0.213
1.01 (0.94, 1.08) Reference
0.815
0.83 (0.71, 0.98) Reference
0.029
0.83 (0.73, 0.93) Reference
0.002
0.86 (0.84, 0.88) Reference
<0.001
0.85 (0.81, 0.90) Reference
<0.001
1.47 (1.30, 1.66) 1.47 (1.31, 1.65) 0.68 (0.58, 0.79) Reference
<0.001 <0.001 <0.001
1.42 (1.38, 1.46) 1.34 (1.30, 1.38) 0.82 (0.79, 0.85) Reference
<0.001 <0.001 <0.001
1.45 (1.37, 1.54) 1.30 (1.23, 1.37) 0.79 (0.73, 0.84) Reference
<0.001 <0.001 <0.001
0.89 (0.78, 1.01) 0.98 (0.89, 1.07) 0.78 (0.71, 0.86) Reference
0.081 0.600 <0.001
1.01 (0.98, 1.03) 1.01 (0.99, 1.03) 0.94 (0.92, 0.96) Reference
0.646 0.447 <0.001
0.81 (0.76, 0.86) 1.04 (1.00, 1.09) 0.88 (0.84, 0.92) Reference
<0.001 0.076 <0.001
Abbreviations: IRR, incidence rate ratio; CI, confidence interval; RV1, Rotarix, GSK; ER, emergency room. a Adjusted for age, calendar year and gender.
younger were significantly more likely than children aged 3 to 4 years to experience RV episodes, RV-related inpatient and ER visits, and diarrhea-related visits of any kind, after adjusting for vaccination status, gender and calendar year. Findings from these multivariate analyses provide further evidence that younger patients are at higher risk for RV-related and diarrhea-related events and thus the associated resource utilization. For acute gastroenteritis hospitalizations among children under five years of age in the pre-vaccination era in the US, approximately 40% occurred in children less than 1 year old, approximately 30% in children 12–23 months and approximately 30% among children 24–59 months [19]. The current study also showed there were significant differences in RV episodes, and RV-related and diarrhea-related events by calendar year after adjusting for vaccination status, gender and age. This is consistent with prior studies which have shown year-to-year variability in acute gastroenteritis and rotavirus [14,21]. The current study found statistically significantly lower rates of RV-related resource utilization and diarrhea-related inpatient and ER visit rates for the complete vaccination versus unvaccinated
cohort, yet no difference for diarrhea-related outpatient visits. Other studies on diarrhea-related resource utilization and RV vaccination have yielded mixed findings regarding outpatient visits as well. A similar analysis to the one described here was conducted with the same data source but using data on patients vaccinated with either RV1 or RV5 [9]. Diarrhea-related outpatient visits were significantly higher among the RV vaccinated versus unvaccinated while the opposite was found for inpatient and ER visits; however, this analysis did not adjust for age and calendar year which may have confounded the associations. The study by Cortes et al. [5], with claims data similar to those in the current study, which examined diarrhea-related resource utilization before and after the introduction of RV5 found marked reductions in diarrhea-related hospitalizations following vaccine introduction but only modest, if any, reductions for emergency department and outpatient visits. In their age- and calendar year-adjusted analysis of vaccinated versus unvaccinated children in the era of vaccine availability, rates of diarrhea-related hospitalization, ER and outpatient visits were significantly reduced but substantially moreso for hospitalization and ER. The authors stated that RV accounts for a smaller proportion of
Please cite this article in press as: Krishnarajah G et al. Public health impact of Rotarix vaccination among commercially insured children in the United States. Vaccine (2017), http://dx.doi.org/10.1016/j.vaccine.2017.06.034
7
G. Krishnarajah et al. / Vaccine xxx (2017) xxx–xxx Table 5 Comparison of mean cost of first RV episode and first diarrhea-related episode between vaccinated and unvaccinated cohorts. Cost per 1000 persons mean ($2012)
Unadjusted difference ($2012) (95% CI)
[A]
[A] – [B]
[B]
p-value
Adjusted differencea ($2012) (95% CI)
p-value
Complete vaccination versus unvaccinated Complete vaccination Unvaccinated First RV episode Total costs Inpatient costs Outpatient costs ER costs
788 378 243 167
First diarrhea episode Total costs 83,203 Inpatient costs 26,348 Outpatient costs 43,140 ER costs 13,714
16,040 11,682 2358 2000
15,252 11,305 2114 ( 1833 (
( 19,492; 11,595) ( 15,046; 8036) 2736; 1600) 2243; 1464)
<0.001 <0.001 <0.001 <0.001
11,511 8767 ( 1381 ( 1364 (
( 12,024; 9855) 9563; 7230) 2052; 972) 1932; 1114)
<0.001 <0.001 <0.001 <0.001
147,033 67,028 54,960 25,045
63,831 40,680 11,820 11,331
( ( ( (
<0.001 <0.001 <0.001 <0.001
46,772 37,246 1848 ( 7679 (
( 66,604; 26,268) ( 61,223; 22,895) 5971; 989) 10,563; 6235)
<0.001 <0.001 0.063 <0.001
81,845; 58,305; 16,262; 13,399;
43,928) 22,838) 5668) 9093)
Incomplete vaccination versus unvaccinated Incomplete vaccination Unvaccinated First RV episode Total costs Inpatient costs Outpatient costs ER costs
2492 1893 313 286
First diarrhea episode Total costs 81,963 Inpatient costs 26,374 Outpatient costs 38,994 ER costs 16,595
16,040 11,682 2358 2000
13,548 9789 ( 2045 ( 1714 (
( 18,796; 7586) 14,554; 4358) 2736; 1387) 2212; 1199)
<0.001 0.006 <0.001 <0.001
9809 7228 1323 1258
( ( ( (
147,033 67,028 54,960 25,045
65,070 40,654 15,967 8450 (
( 91,085; 31,167) ( 63,508; 11,511) ( 22,849; 7081) 12,428; 4511)
<0.001 0.014 0.006 0.004
47,703 36,795 5971 ( 4938 (
11,511; 5413) 9355; 2692) 2145; 979) 1364; 235) ( 79,530; 39,433) ( 47,246; 7679) 8900; 1114) 6684; 1232)
<0.001 <0.001 <0.001 <0.001 <0.001 0.004 <0.001 <0.001
Abbreviations: CI, confidence interval; ER, emergency room. a Adjusted for birth year and gender.
diarrhea events in outpatient and ER settings than hospitals. This may render findings based on all diarrhea-related events less informative in some settings, such as outpatient offices. More recently, a study by Leshem et al. which adjusted for age and calendar year reported that outpatient visits for diarrhea among either RV1 or RV5 vaccinated individuals and unvaccinated individuals during the 2010–2011 RV season were similar [10]. Furthermore, a study by Cortese et al. evaluated protection against gastroenteritis among household members where children were vaccinated against RV [20]. In this age-adjusted analysis, outpatient gastroenteritis encounters were modestly increased, whereas inpatient and emergency department visits for gastroenteritis were either decreased or unchanged for household members with an RV vaccinated child versus for household members with an unvaccinated child [20]. The authors speculated there was lower wild-type RV shedding from a vaccinated versus unvaccinated child, and therefore less severe RV cases among other household members of vaccinated children which might lead to more outpatient visits and not acute care visits. Furthermore, in clinical trials, RV vaccination was very effective against hospitalization but far less with mild infections [21]. It is also plausible that vaccinated and unvaccinated patients may seek care for diarrheal diseases in different settings. The economic analysis of RV burden also demonstrated lower cost burden for both complete and incomplete RV1 vaccination relative to unvaccinated children. In the current study’s adjusted analysis, the difference in mean cost for first RV episode was $11.51 for unvaccinated versus completely vaccinated children. Thus, if each of the 34,928 children in the current study who had been vaccinated had not been vaccinated, the cost of first RV episode would have been $402,056 greater for this cohort. There are some limitations of the current study. First, we examined data from only four post-RV1 vaccine seasons (2008–2011), and there were few children in the vaccinated cohorts in the first season. Delayed uptake of RV1 vaccination resulted in follow-up of the vaccinated cohorts having a higher proportion of younger
children with shorter follow-up time relative to the unvaccinated cohort. We cannot be certain that observed changes in the postRV vaccine years were due solely to vaccine use. Secular trends in incidence of RV and other diarrheal pathogens could affect our findings. In addition, unvaccinated children in this study gained herd immunity from those vaccinated with either RV1 or RV5 as both vaccines were available; thus it is difficult to measure the actual impact of RV1 as the unvaccinated cohort receives benefit of immunity from both RV1 and RV5 vaccinated individuals. ICD9 codes were used to identify diagnoses; these codes may not reflect confirmed clinical diagnoses and lack information to assess illness severity. Moreover, medical services obtained outside a patient’s insurance plan are not captured in a claims database. Specifically, RV vaccination obtained outside one’s health plan could not be ascertained and could have resulted in some individuals being misclassified as unvaccinated; this would have biased the results toward the null. There is also potential for underestimating the resource utilization and costs. Finally, inability to adjust for socioeconomic status is a limitation. Given the effectiveness of RV1 vaccination and RV vaccination in general, efforts should focus on improving RV vaccination rates which lag behind most pediatric vaccination rates and continue to fall below the Healthy People 2020 target of 80% [18].
Disclosure statement GK was employed by the GSK group of companies at the time of the study conduct and during the development of the manuscript and is currently employed by CSL Behring. GK also reports ownership of stock options/restricted shares from the GSK group of companies and CSL Behring. AK, CK, MSD, and PL are employees of Analysis Group which has received funding from the GSK group of companies to conduct this study and has received funding for other drug safety and health outcomes research studies by multiple pharmaceutical companies.
Please cite this article in press as: Krishnarajah G et al. Public health impact of Rotarix vaccination among commercially insured children in the United States. Vaccine (2017), http://dx.doi.org/10.1016/j.vaccine.2017.06.034
8
G. Krishnarajah et al. / Vaccine xxx (2017) xxx–xxx
Role of the funding source GlaxoSmithKline Biologicals SA funded this study and was involved in all stages of study conduct, including study design; collection, analysis, and interpretation of the data; writing the report; and the decision to submit the paper for publication. GlaxoSmithKline Biologicals SA also paid all costs associated with the development and publication of this manuscript. Contributions All authors participated in the design or implementation or analysis, and interpretation of the study; and the development of this manuscript. All authors had full access to the data and gave final approval before submission. All authors have agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. Trademark statement Rotarix is a trademark of the GSK group of companies. RotaTeq is a trademark of Merck & Co., Inc. Acknowledgement The authors thank Ning Wu (freelance on behalf of GSK) for writing support and editorial assistance. The authors would like to thank Business & Decision Life Sciences platform for editorial assistance and manuscript coordination, on behalf of GSK. Marie Cloes coordinated manuscript development and provided editorial support. Heather Santiago (publication manager, GSK) also contributed to manuscript coordination and editorial assistance. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.vaccine.2017.06. 034. References [1] Cortese MM, Parashar UD. Centers for Disease Control and Prevention. Prevention of rotavirus gastroenteritis among infants and children: recommendations of the Advisory Committee on Immunization Practices (ACIP). MMWR Recomm Rep 2009;58:1–25. [2] Parashar UD, Alexander JP, Glass RI. Advisory Committee on Immunization Practices Centers for Disease Control and Prevention. Prevention of rotavirus gastroenteritis among infants and children. Recommendations of the Advisory Committee on Immunization Practices (ACIP). MMWR Recomm Rep 2006;55:1–13.
[3] Highlights of prescribing information for Rotarix. Available at: [accessed December 1, 2016]. [4] Velazquez RF, Linhares AC, Munoz S, Seron P, Lorca P, DeAntonio R, OrtegaBarria E. Efficacy, safety and effectiveness of licensed rotavirus vaccines: a systematic review and meta-analysis for Latin America and the Carribean. BMC Pediatrics 2017;17:14. [5] Cortes JE, Curns AT, Tate JE, Cortese MM, Patel MM, Zhou F, et al. Rotavirus vaccine and health care utilization for diarrhea in US children. N Engl J Med 2011;365:1108–17. [6] Wang FT, Mast TC, Glass RJ, Loughlin J, Seeger JD. Effectiveness of the pentavalent rotavirus vaccine in preventing gastroenteritis in the United States. Pediatrics 2010;125:e208–13. [7] Boom JA, Tate JE, Sahni LC, Rench MA, Hull JJ, Gentsch JR, et al. Effectiveness of pentavalent rotavirus vaccine in a large urban population in the United States. Pediatrics 2010;125:e199–207. [8] Payne DC, Boom JA, Staat MA, Edwards KM, Szilagyi PG, Klein EJ, et al. Effectiveness of pentavalent and monovalent rotavirus vaccines in concurrent use among US children <5 years of age, 2009–2011. Clin Infect Dis 2013;57:13–20. [9] Krishnarajah G, Duh M, Korves C, Demissie K. Public health impact of complete and incomplete rotavirus vaccination among commercially and Medicaid insured children in the United States. PLOS ONE 2016. DOI:10.1371/journal. pone.0145977. [10] Leshem E, Moritz RE, Curns AT, Zhou F, Tate JE, Lopman BA, et al. Rotavirus vaccines and health care utilization for diarrhea in the United States (2007– 2011). Pediatrics 2014;134:15–23. [11] Krishnarajah G, Demissie K, Lefebvre P, Gaur S, Duh MS. Clinical and cost burden of rotavirus infection before and after introduction of rotavirus vaccines among commercially and Medicaid insured children in the United States. Hum Vaccin Immunother 2014;10:2255–66. [12] American College of Physicians Understanding Capitation. Available at: [accessed May 2, 2017]. [13] Hsu VP, Staat MA, Roberts N, Thieman C, Bernstein DI, Bresee J, et al. Use of active surveillance to validate international classification of diseases code estimates of rotavirus hospitalizations in children. Pediatrics 2005;115:78–82. [14] Curns AT, Coffin F, Glasser JW, Glass RI, Parashar UD. Projected impact of the new rotavirus vaccination program on hospitalizations for gastroenteritis and rotavirus disease among US children <5 years of age during 2006–2015. J Infect Dis 2009;200(Suppl 1):S49–56. [15] Barker N. A practical introduction to the bootstrap using the SAS system. Available at: [accessed January 7, 2014]. [16] Wang FT, Mast TC, Glass RJ, Loughlin J, Seeger JD. Effectiveness of an incomplete RotaTeq (RV5) vaccination regimen in preventing rotavirus gastroenteritis in the United States. Pediatr Infect Dis J 2013;32:278–83. [17] Centers for Disease Control and Prevention. National, state, and local area vaccination coverage among children aged 19–35 months – United States, 2011. MMWR 2012;61(35):689–96. [18] Hill HA, Elam-Evans LD, Yankey D, Singleton JA, Dietz V. Vaccination coverage among children aged 19–35 months – United States, 2015. MMWR 2016;65 (39):1065–71. [19] Parashar UD, Holman RC, Clarke MJ, Bresee JS, Glass RI. Hospitalizations associated with rotavirus diarrhea in the United States, 1993 through 1995: surveillance based on the new ICD-9-CM rotavirus-specific diagnosis code. J Infect Dis 1998;177:13–7. [20] Cortese MM, Dahl RM, Curns AT, Parashar UD. Protection against gastroenteritis in US households with children who received rotavirus vaccine. J Infect Dis 2015;211:558–62. [21] Anderson EJ. Time to begin a new chapter and expand rotavirus immunization. Clin Infect Dis 2014;59:982–6.
Please cite this article in press as: Krishnarajah G et al. Public health impact of Rotarix vaccination among commercially insured children in the United States. Vaccine (2017), http://dx.doi.org/10.1016/j.vaccine.2017.06.034
Contents lists available at ScienceDirect
Vaccine journal homepage: www.elsevier.com/locate/vaccine
Public health impact of Rotarix vaccination among commercially insured children in the United States Girishanthy Krishnarajah a,1, Andrew Kageleiry b, Caroline Korves b, Patrick Lefebvre c, Mei S. Duh b,⇑ a
GSK, 5 Crescent Drive, Philadephia, PA 19112 United States Analysis Group, Inc., 111 Huntington Avenue, 14th Floor, Boston, MA 02199, United States c Groupe d’analyse, Ltée, 1000 De La Gauchetière Ouest, Bureau 1200, Montréal, Quebec H3B 4W5, Canada b
a r t i c l e
i n f o
Article history: Received 16 February 2017 Received in revised form 25 May 2017 Accepted 10 June 2017 Available online xxxx Keywords: Rotavirus Rotarix Diarrhea Gastroenteritis Commercially insured
a b s t r a c t Background: This study (NCT01915888) assessed public health impact of Rotarix, GSK [RV1] vaccination. Methods: Children born between 2007–2011 were identified from Truven Commercial Claims and Encounters Databases and observed until earlier of plan disenrollment or five years old. Children receiving one or two doses of RV1 during the vaccination window were assigned to incomplete and complete vaccination cohorts, respectively. Children without rotavirus (RV) vaccination (RV1 OR RotaTeq, Merck & Co., Inc. [RV5]) were assigned to the unvaccinated cohort. Claims with International Classification of Disease 9th edition (ICD-9) codes for diarrhea and RV infections were identified. First RV episode incidence, RV-related and diarrhea-related healthcare resource utilization were compared. Multivariate Poisson regression with generalized estimating equations was used to generate 95% confidence intervals (CIs) around incidence rate ratios (IRR) between cohorts while adjusting for gender, age and calendar year. Mean costs for first RV and diarrhea episodes were calculated with adjustment for gender and birth year; bootstrapping was used to determine statistically significant differences between cohorts. Results: Incidence of first RV episodes was significantly reduced in complete and incomplete vaccination cohorts compared to the unvaccinated cohort (IRR = 0.17 [95%CI: 0.09–0.30] and IRR = 0.19 [95%CI: 0.06– 0.58], respectively). RV-related inpatient, outpatient and emergency room (ER) visits were significantly lower for complete vaccination versus unvaccinated cohort. Diarrhea-related inpatient and ER visit rates were significantly lower for complete vaccination versus unvaccinated cohorts; outpatient rates were similar. RV-related and diarrhea-related resource utilization rates were significantly lower or no different for incomplete vaccination versus unvaccinated cohort. Compared with unvaccinated children, adjusted mean cost for first RV episode and first diarrhea episode per 1000 persons was $11,511 (95%CI: $9855$12,024) and $46,772 (95%CI: $26,268-$66,604) lower, respectively, for completely vaccinated children. Conclusions: RV1 vaccination confers benefits in reduction of RV incidence, RV- and diarrhea-related healthcare resource utilization, and RV- and diarrhea-related healthcare costs. Ó 2017 GlaxoSmithKline. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
1. Introduction Before initiation of a routine infant rotavirus (RV) vaccination program in February 2006 in the United States (US), nearly every Abbreviations: RV, rotavirus; RV1, Rotarix, GSK; RV5, RotaTeq, Merck & Co., Inc.; ICD-9, International Classification of Disease 9th edition; CI, confidence interval; ACIP, Advisory Committee on Immunization Practices; ER, emergency room; IR, incidence rate; IRR, incidence rate ratio; PI, package insert. ⇑ Corresponding author. E-mail addresses: [email protected] (G. Krishnarajah), Andrew. [email protected], [email protected] (A. Kageleiry), Caroline. [email protected] (C. Korves), [email protected] (P. Lefebvre), [email protected] (M.S. Duh). 1 Present address: CSL Behring, 1020 1st Avenue, King of Prussia, Pennsylvania 19406 United States.
child was infected with RV by age 5 years [1]. Since then, both the Advisory Committee on Immunization Practices (ACIP) and the American Academy of Pediatrics have recommended routine immunization of children with RV vaccine [1,2]. Two brands of oral RV vaccines are approved for children in the US by its Food and Drug Administration: RotaTeq, Merck & Co., Inc. (RV5) (approved in 2006) and Rotarix, GSK (RV1) (approved in 2008) [1]. RV1 is administered in two doses; the vaccine is administered in three doses if RV5 or a mix of brands (i.e., both RV1 and RV5) is used. The ACIP recommends a vaccination window from 6 weeks old to 8 months old [1]. The package insert (PI) for RV1 specifies the first of two doses should be administered to children beginning at 6 weeks of age, and the second should be administered with at least a four week gap from the first dose and before 24 weeks of age [3].
http://dx.doi.org/10.1016/j.vaccine.2017.06.034 0264-410X/Ó 2017 GlaxoSmithKline. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
Please cite this article in press as: Krishnarajah G et al. Public health impact of Rotarix vaccination among commercially insured children in the United States. Vaccine (2017), http://dx.doi.org/10.1016/j.vaccine.2017.06.034
2
G. Krishnarajah et al. / Vaccine xxx (2017) xxx–xxx
Payers need to evaluate the impact of RV vaccination on both clinical and economic endpoints from a real-world perspective. RV1 and RV5, with their different antigen compositions and dosing schedules [4], are both currently in use in the US. Published studies to date using real-world practice data in the US to evaluate effectiveness of complete and incomplete vaccination have focused only on RV5 [5–7] or both RV1 and RV5 [8–10]. Studies on vaccine effectiveness in real world settings provide important public health data and can contribute to decision-making about ongoing vaccination programs. The present study attempts to address current gaps of knowledge by estimating the impact of RV1 vaccination on RV incidence and healthcare resource utilization among children aged less than 5 years using commercial insurance claims data. 2. Methods 2.1. Data source Data from the 2007–2011 Truven Commercial Claims and Encounters database were analyzed. Data from the 2007–2011 Truven Medicaid database were analyzed, but the sample size was inadequate for valid statistical comparisons as the cohorts for complete vaccination and incomplete vaccination were approximately 2000, and there was only one RV episode per each cohort; these results are not presented. Truven databases are described in detail in Krishnarajah et al. [11]. The current study (NCT01915888) utilized 2007–2011 data captured from approximately 9 million children under 5 years of age with commercial insurance coverage. 2.2. Patient selection Children were further required to have continuously received medical and pharmacy benefits since birth until end of follow-up (defined as disenrollment from insurance plans, 5 years of age, death, or data cutoff, whichever occurred first) and not to have enrolled in capitation-based health plans. With capitation-based health plans, participating physicians are paid a pre-determined amount of money per patient enrolled for a specified range of services [12]. As there is no fee for these services, administrative databases may have incomplete records for utilization and cost. Children residing in states with universal vaccination programs were excluded from the analysis as done in prior studies [10]; vaccinations in these states were not likely billed to insurance companies; thus, no record of vaccination would appear in the database. Children receiving any RV5 vaccination during follow-up were excluded. 2.3. Study cohorts Children were assigned to cohorts based on RV1 vaccines received. Individuals’ person-time was assessed from 6 weeks through 24 weeks old to determine RV1 vaccination exposure. RV1 vaccination was identified by current procedural terminology code 90681 or national drug code 58160-805-01, 58160-805-02, 58160-805-11, 58160-851-01, 58160-853-02, and 58160-854-52. As insurance claims data do not include date of birth, the authors provided Truven Health Analytics, Inc. with dates of service for RV1 vaccination. Truven Health Analytics, Inc., with access to exact date of birth data, identified whether vaccination was given during the 6–24 week age window. Children given two RV1 vaccinations during the vaccination window at least four weeks apart contributed person-time to the complete vaccination cohort; those with one RV1 vaccination contributed to the incomplete vaccination cohort; children without vaccination contributed to the
unvaccinated cohort. The unvaccinated cohort was comprised of individuals who reached 24 weeks of age old on or after January 1, 2008; originally, January 1, 2006 was the cut-off date, but this was changed as only RV5 (and not RV1) was available in the 2006–2008 timeframe, and non-vaccination during the time of RV1 availability was of interest. 2.4. Outcomes Incidence rate ratios (IRR) of the first RV episode were estimated to compare rates between cohorts. The first occurrence of an RV-coded encounter marked incidence of first RV episode. RVcoded encounters were identified by claims with the specific International Classification of Diseases 9th Edition (ICD-9) code for RV (008.61 - Enteritis due to rotavirus). A single RV episode included groups of RV-related medical claims dated less than 28 days apart. Children who had the first RV episode prior to 24 weeks of age did not contribute to the RV incidence estimation. The rate of RV-related healthcare resource utilization was calculated with the numerator being the number of RV-related hospitalizations, outpatient visits or emergency room (ER) visits (including visits to ER or urgent care facility) occurring on different calendar days, and the denominator being the total follow-up time since end of vaccination window within a cohort. Costs per first RV episode were calculated as the sum of all medical claims dated within the 14 days before the first RV-coded claim and the 14 days after the last RV-coded claim within an RV episode where any RVcoded claims occurring within a 14 day period were considered to be the same episode [9,11]. An individual without an RV episode thus had a cost of $0. Because laboratory tests confirming RV infection are not performed routinely on all patients with diarrhea, claims coded as RV likely represent only a fraction of all RV-related encounters [5,13]; therefore, diarrhea incidence was also estimated. Diarrhea-coded healthcare encounters were identified by claims with non-specific ICD-9 codes for diarrhea. See Supplementary Table 1 for details. Similar methods were used to calculate rate of diarrhea-related visits and costs for first diarrhea episodes. 2.5. Statistical analysis Analyses were conducted according to PI recommendations, and frequencies and proportions were reported to describe baseline characteristics of the study cohorts. Pearson’s chi-square test was used to compare differences between cohorts. Poisson regression models with generalized estimating equations were used to estimate IRRs and 95% confidence intervals (CIs) for RV- and diarrhea-related outcomes. Both univariate Poisson regressions and post hoc multivariate Poisson regressions were performed. Since the likelihood of RV infection varies with age and calendar year [14], in part due to the change in prevalence of the virus which is affected by vaccination, multivariate analyses were necessary to account for potential confounding by these factors as they were significantly different between cohorts. Mean cost per 1000 children per first RV episode was calculated for each cohort by summing episode costs for children within a cohort and dividing that by the number of children within the cohort, and then multiplying that by 1000. The difference in mean cost per 1000 children per first RV episode was then calculated by subtracting the mean of one cohort from that of another [9,11]. Univariate and post hoc multivariate linear regression (with adjustment for gender and birth year) were performed. Cost data are often not normally distributed [15]. Thus, p-values and 95% CIs for cost differences between vaccination cohorts were computed using bootstrapping (with replacement) with 999 iterations. Similarly, diarrhearelated resource use and costs of the first diarrhea episode were
Please cite this article in press as: Krishnarajah G et al. Public health impact of Rotarix vaccination among commercially insured children in the United States. Vaccine (2017), http://dx.doi.org/10.1016/j.vaccine.2017.06.034
3
G. Krishnarajah et al. / Vaccine xxx (2017) xxx–xxx
estimated and compared between cohorts. Costs were inflationadjusted to 2012 US dollars based on the medical care component of the Consumer Price Index. All analyses were performed using SAS software version 9.3 (SAS Institute, Inc., Cary, NC, USA).
3. Results A total of 225,587 children born between 2007 and 2011 with commercial insurance were eligible for the analysis (Table 1). Of
Table 1 Characteristics of eligible children. Complete vaccination [A]
Incomplete vaccination [B]
Unvaccinated [C]
p-values A vs. B
A vs. C
B vs. C
Overall study period (2008–2011)
N = 34,928
N = 8390
N = 182,269
Year of birth, n (%) 2008 2009 2010 2011
3392 (9.7%) 17,574 (50.3%) 13,571 (38.9%) 391 (1.1%)
913 (10.9%) 4146 (49.4%) 3246 (38.7%) 85 (1.0%)
65,465 (35.9%) 49,655 (27.2%) 39,934 (21.9%) 655 (0.4%)
0.001 0.139 0.780 0.401
<0.001 <0.001 <0.001 <0.001
<0.001 <0.001 <0.001 <0.001
Gender, n (%) Female
16,871 (48.3%)
4054 (48.3%)
88,137 (48.4%)
0.977
0.855
0.949
Type of health plan, n (%) Comprehensive Preferred provider organization Other
187 (0.5%) 26,514 (75.9%) 8227 (23.6%)
49 (0.6%) 6314 (75.3%) 2027 (24.2%)
1969 (1.1%) 140,502 (77.1%) 39,798 (21.8%)
0.587 0.209 0.241
<0.001 <0.001 <0.001
<0.001 <0.001 <0.001
Region, n (%) Northeast North Central South West Unknown
4819 (13.8%) 10,310 (29.5%) 14,902 (42.7%) 4320 (12.4%) 577 (1.7%)
1199 (14.3%) 2118 (25.2%) 3793 (45.2%) 1162 (13.8%) 118 (1.4%)
22,110 (12.1%) 57,016 (31.3%) 72,823 (40.0%) 27,018 (14.8%) 3302 (1.8%)
0.240 <0.001 <0.001 <0.001 0.108
<0.001 <0.001 <0.001 <0.001 0.039
<0.001 <0.001 <0.001 0.014 0.006
9173 (26.3%)
2213 (26.4%)
40,471 (22.2%)
0.831
<0.001
<0.001
4493 3139 3589 3778
1208 (14.4%) 865 (10.3%) 948 (11.3%) 1025 (12.2%)
23,079 15,317 16,459 18,352
<0.001 <0.001 0.006 <0.001
0.300 <0.001 <0.001 <0.001
<0.001 <0.001 <0.001 <0.001
3542 (10.1%) 2110 (6.0%)
853 (10.2%) 540 (6.4%)
11,722 (6.4%) 10,250 (5.6%)
0.944 0.175
<0.001 0.002
<0.001 0.002
1908 (5.5%) 985 (2.8%)
531 (6.3%) 287 (3.4%)
9758 (5.4%) 5990 (3.3%)
0.002 0.003
0.408 <0.001
<0.001 0.500
1799 (5.2%)
447 (5.3%)
7493 (4.1%)
0.511
<0.001
<0.001
Year 2008 Age at start of year, n (%) Born in year 0–1 years 1–2 years 2–3 years 3–4 years
N = 181
N = 123
N = 61,719
181 (100.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%)
123 (100.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%)
35,159 (57.0%) 26,560 (43.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%)
1.000 1.000 1.000 1.000 1.000
<0.001 <0.001 1.000 1.000 1.000
<0.001 <0.001 1.000 1.000 1.000
Year 2009 Age at start of year, n (%) Born in year 0–1 years 1–2 years 2–3 years 3–4 years
N = 14,373
N = 3132
N = 107,979
11,006 (76.6%) 3367 (23.4%) 0 (0.0%) 0 (0.0%) 0 (0.0%)
2238 (71.5%) 894 (28.5%) 0 (0.0%) 0 (0.0%) 0 (0.0%)
29,049 (26.9%) 59,521 (55.1%) 19,409 (18.0%) 0 (0.0%) 0 (0.0%)
<0.001 <0.001 1.000 1.000 1.000
<0.001 <0.001 <0.001 1.000 1.000
<0.001 <0.001 <0.001 1.000 1.000
Year 2010 Age at start of year, n (%) Born in year 0–1 years 1–2 years 2–3 years 3–4 years
N = 22,494
N = 5577
N = 107,971
6256 (27.8%) 14,086 (62.6%) 2152 (9.6%) 0 (0.0%) 0 (0.0%)
1632 (29.3%) 3396 (60.9%) 549 (9.8%) 0 (0.0%) 0 (0.0%)
22,451 (20.8%) 39,014 (36.1%) 34,763 (32.2%) 11,743 (10.9%) 0 (0.0%)
0.031 0.017 0.530 1.000 1.000
<0.001 <0.001 <0.001 <0.001 1.000
<0.001 <0.001 <0.001 <0.001 1.000
Year 2011 Age at start of year, n (%) Born in year 0–1 years 1–2 years 2–3 years 3–4 years
N = 23,987
N = 5668
N = 92,417
391 (1.6%) 11,973 (49.9%) 10,076 (42.0%) 1547 (6.4%) 0 (0.0%)
85 (1.5%) 2834 (50.0%) 2369 (41.8%) 380 (6.7%) 0 (0.0%)
655 (0.7%) 33,340 (36.1%) 26,236 (28.4%) 23,873 (25.8%) 8313 (9.0%)
0.482 0.908 0.773 0.484 1.000
<0.001 <0.001 <0.001 <0.001 <0.001
<0.001 <0.001 <0.001 <0.001 <0.001
Comorbidities, n (%) Acute upper respiratory infections of multiple or unspecified sites Suppurative and unspecified otitis media Acute bronchitis and bronchiolitis General symptoms Symptoms involving respiratory system and other chest symptoms Diseases of esophagus Viral and chlamydial infection in conditions classified elsewhere and of unspecified site Symptoms involving digestive system Nonsuppurative otitis media and Eustachian tube disorders Atopic dermatitis and related conditions
(12.9%) (9.0%) (10.3%) (10.8%)
(12.7%) (8.4%) (9.0%) (10.1%)
Please cite this article in press as: Krishnarajah G et al. Public health impact of Rotarix vaccination among commercially insured children in the United States. Vaccine (2017), http://dx.doi.org/10.1016/j.vaccine.2017.06.034
4
G. Krishnarajah et al. / Vaccine xxx (2017) xxx–xxx
those who reached 24 weeks of age on or after January 1, 2008, 182,269 (80.8%) served as the unvaccinated cohort; 34,928 (15.5%) were classified as completely vaccinated; the remaining 8390 (3.7%) were grouped as the incomplete vaccination cohort. Year of birth varied significantly across cohorts (Table 1). In the unvaccinated cohort, 35.9% of individuals were born in 2008 whereas approximately 10% in the vaccination cohorts were born that year (p-value <0.001 for complete vaccination versus unvaccinated, and for incomplete vaccination versus unvaccinated). The proportion of patients in the vaccinated cohorts born during 2009 and 2010 was nearly double that of the unvaccinated cohort. Thus, for this study, the vaccinated cohorts had more observation time among younger children and in later calendar years relative to the unvaccinated cohort. Following the vaccination window, there were 11, 3 and 449 RV episodes observed in the complete vaccination, incomplete vaccination and unvaccinated cohorts, respectively (Supplementary Table 2). Using the multivariate model adjusting for gender, age and calendar year, the IRR for RV episodes for the complete vaccination cohort and incomplete vaccination cohorts versus the unvaccinated cohort were 0.17 (95% CI: 0.09–0.30) and 0.19 (95% CI: 0.06–0.58), respectively (Table 2). In the multivariate analysis of RV-related resource utilization (Supplementary Table 3 and Table 3), complete vaccination versus no vaccination was associated with significant reductions in inpatient visits (IRR = 0.07 [95% CI: 0.02–0.21]), outpatient visits (IRR = 0.25 [95% CI: 0.11–0.54]), and ER visits (IRR = 0.09 [95% CI: 0.03–0.32]). Compared with the unvaccinated cohort, children
with incomplete vaccination also had lower rates of RV-related inpatient (IRR = 0.09 [95% CI: 0.01–0.67]), and ER visits (IRR = 0.19 [95% CI: 0.05–0.78]); there was no significant difference for outpatient visits (IRR = 0.41 [95% CI: 0.09–1.94]). When comparing rates of diarrhea-related medical encounters between children with complete vaccination versus no vaccination, those completely vaccinated had statistically significantly lower rates of inpatient visits (IRR = 0.37 [95% CI: 0.29–0.47]), and ER visits (IRR = 0.62 [95% CI: 0.57–0.68]); there was no significant difference in outpatient visits (IRR = 1.02 [95% CI: 0.98–1.06]) (Supplementary Table 4 and Table 4). Compared with the unvaccinated cohort, children with incomplete vaccination had similar rates of diarrhea-related inpatient (IRR = 0.78 [95% CI: 0.52– 1.16]) and outpatient visits (IRR = 1.01 [95% CI: 0.94–1.08]). Incompletely vaccinated children had a statistically significantly lower rate of ER visits relative to the unvaccinated (IRR = 0.83 [95% CI: 0.71–0.98]) (Supplementary Table 4 and Table 4). Adjusted mean total healthcare cost per 1000 persons for first RV episode in the complete vaccination cohort was significantly lower than that in the unvaccinated cohort (difference = $11,511 [95% CI: $12,024 to $9855]) (Table 5). Similar cost difference was observed between the incomplete vaccination and unvaccinated cohorts (difference = $9809 [95% CI: $11,511 to $5413]). Across comparisons, differences were largely driven by statistically significant differences in inpatient visit costs. The mean costs of individual first diarrhea episodes per 1000 persons were much higher than the mean costs of the first RV episodes per 1000 persons, mainly because of the much higher incidence
Table 2 Comparison of incidence of first RV episode between vaccinated and unvaccinated cohorts. p-value
Adjusted IRR (95% CI)a
p-value
<0.001
0.17 (0.09, 0.30) Reference
<0.001
Gender Female Male
0.80 (0.67, 0.96) Reference
0.019
Age (years) 0–1 1–2 2–3 3–4
1.45 (1.18, 1.77) 1.52 (1.25, 1.85) 0.73 (0.56, 0.95) Reference
<0.001 <0.001 0.022
Year 2008 2009 2010 2011
0.99 (0.80, 1.23) 1.25 (1.08, 1.44) 0.39 (0.32, 0.48) Reference
0.920 0.003 <0.001
0.19 (0.06, 0.58) Reference
0.004
Gender Female Male
0.81 (0.67, 0.97) Reference
0.023
Age (years) 0–1 1–2 2–3 3–4
1.47 (1.20, 1.80) 1.55 (1.28, 1.89) 0.71 (0.54, 0.93) Reference
<0.001 <0.001 0.012
Year 2008 2009 2010 2011
0.98 (0.79, 1.22) 1.23 (1.06, 1.42) 0.40 (0.32, 0.49) Reference
0.878 0.005 <0.001
Unadjusted IRR (95% CI) Complete vaccination versus unvaccinated RV1 vaccination cohort Complete vaccination 0.18 (0.10, 0.33) Unvaccinated Reference
Incomplete vaccination versus unvaccinated RV1 vaccination cohort Incomplete vaccination 0.21 (0.07, 0.64) Unvaccinated Reference
0.006
Abbreviations: IRR, incidence rate ratio; CI, confidence interval; RV, rotavirus; RV1, Rotarix, GSK. a Adjusted for age, calendar year and gender.
Please cite this article in press as: Krishnarajah G et al. Public health impact of Rotarix vaccination among commercially insured children in the United States. Vaccine (2017), http://dx.doi.org/10.1016/j.vaccine.2017.06.034
5
G. Krishnarajah et al. / Vaccine xxx (2017) xxx–xxx Table 3 Comparison of RV-related resource utilization between vaccinated and unvaccinated cohorts. IRR (95% CI)
p-value
IRR (95% CI)
p-value
Complete vaccination versus unvaccinated Inpatient visits Outpatient visits Unadjusted RV1 vaccination cohort Complete vaccination Unvaccinated Adjusteda RV1 vaccination cohort Complete vaccination Unvaccinated Gender Female Male Age (years) 0–1 1–2 2–3 3–4 Year 2008 2009 2010 2011
Adjusteda RV1 vaccination cohort Incomplete vaccination Unvaccinated Gender Female Male Age (years) 0–1 1–2 2–3 3–4 Year 2008 2009 2010 2011
p-value
ER visits
0.08 (0.03, 0.26) Reference
<0.001
0.25 (0.13, 0.47) Reference
<0.001
0.11 (0.04, 0.31) Reference
<0.001
0.07 (0.02, 0.21) Reference
<0.001
0.25 (0.11, 0.54) Reference
<0.001
0.09 (0.03, 0.32) Reference
<0.001
0.90 (0.71, 1.15) Reference
0.412
0.66 (0.49, 0.89) Reference
0.007
0.86 (0.65, 1.15) Reference
0.316
1.59 (1.20, 2.11) 1.93 (1.47, 2.53) 0.69 (0.48, 1.01) Reference
0.001 <0.001 0.054
1.19 (0.87, 1.63) 1.12 (0.84, 1.50) 0.74 (0.50, 1.10) Reference
0.266 0.445 0.138
1.34 (1.00, 1.79) 1.50 (1.16, 1.94) 0.64 (0.44, 0.94) Reference
0.051 0.002 0.023
0.87 (0.65, 1.17) 1.10 (0.91, 1.33) 0.43 (0.33, 0.55) Reference
0.359 0.335 <0.001
1.02 (0.74, 1.41) 1.28 (1.02, 1.60) 0.43 (0.32, 0.59) Reference
0.901 0.036 <0.001
0.74 (0.49, 1.11) 1.21 (0.95, 1.54) 0.41 (0.30, 0.57) Reference
0.143 0.121 <0.001
Incomplete vaccination versus unvaccinated Inpatient visits Outpatient visits Unadjusted RV1 vaccination cohort Incomplete vaccination Unvaccinated
IRR (95% CI)
ER visits
0.12 (0.02, 0.82) Reference
0.031
0.42 (0.16, 1.13) Reference
0.086
0.24 (0.06, 0.97) Reference
0.045
0.09 (0.01, 0.67) Reference
0.018
0.41 (0.09, 1.94) Reference
0.260
0.19 (0.05, 0.78) Reference
0.021
0.92 (0.72, 1.17) Reference
0.480
0.64 (0.47, 0.88) Reference
0.005
0.88 (0.66, 1.17) Reference
0.365
1.58 (1.19, 2.10) 1.95 (1.48, 2.55) 0.70 (0.48, 1.01) Reference
0.002 <0.001 0.056
1.22 (0.89, 1.67) 1.16 (0.86, 1.55) 0.71 (0.48, 1.06) Reference
0.210 0.337 0.095
1.34 (1.00, 1.80) 1.55 (1.19, 2.00) 0.62 (0.42, 0.92) Reference
0.048 <0.001 0.016
0.87 (0.65, 1.17) 1.09 (0.90, 1.32) 0.43 (0.33, 0.55) Reference
0.373 0.387 <0.001
1.01 (0.73, 1.39) 1.25 (0.99, 1.57) 0.44 (0.32, 0.60) Reference
0.952 0.059 <0.001
0.74 (0.50, 1.11) 1.18 (0.93, 1.50) 0.41 (0.30, 0.57) Reference
0.143 0.173 <0.001
Abbreviations: IRR, incidence rate ratio; CI, confidence interval; RV1, Rotarix, GSK; ER, emergency room. a Adjusted for age, calendar year and gender.
of diarrhea-related episodes relative to RV episodes. Across inpatient, outpatient, and ER visit settings, mean costs for first RV episode were uniformly statistically significantly lower for the complete and incomplete vaccination cohorts compared with the unvaccinated cohort. 4. Discussion Findings from the current study suggest that vaccination with RV1 results in significant reduction in RV-coded visits for children following the vaccination window. Higher RV-related healthcare resource utilization and costs were also observed in the unvaccinated cohort versus children with complete RV1 vaccination. Findings from this current study are in agreement with other published studies on RV vaccination with RV5 or mixed vaccination [5–10]. Also compatible with previous findings, this study found lower RV-coded encounters among individuals with incomplete RV vaccination versus those in the unvaccinated cohort [9,16]. A prior study of children under 5 years of age with commercial and Medicaid insurance showed incomplete vaccination with either RV1 or
RV5 versus no vaccination was associated with statistically significant reductions in first RV episodes and RV-related healthcare utilization [9]. A study of RV5 exclusively showed incomplete vaccination versus no vaccination was associated with reductions in both RV-related and diarrhea-related hospitalizations and ER visits [16]. Notably, all of these studies are observational and have relied on insurance claims data. The current study was not designed to evaluate incomplete RV1 vaccination, and incomplete RV1 vaccination has not been evaluated in clinical trials and is not recommended by ACIP. In the current analysis, the complete and incomplete vaccination cohorts were significantly younger than the unvaccinated cohort; approximately 40% of children included in the vaccinated cohorts were born in 2010 and 2011 versus only 22% in the unvaccinated cohort. This observation is consistent with reports that uptake of RV vaccination in general improved in the first several years following availability; nationally, RV vaccination coverage among 19–35 month olds in the US rose from 43.9% in 2009 to 67.3% in 2011 [17] and was reported to be 73.2% in 2015 [18]. In the current study, children aged 2 years and
Please cite this article in press as: Krishnarajah G et al. Public health impact of Rotarix vaccination among commercially insured children in the United States. Vaccine (2017), http://dx.doi.org/10.1016/j.vaccine.2017.06.034
6
G. Krishnarajah et al. / Vaccine xxx (2017) xxx–xxx
Table 4 Comparison of diarrhea-related resource utilization between vaccinated and unvaccinated cohorts. IRR (95% CI)
p-value
IRR (95% CI)
p-value
Complete vaccination versus unvaccinated Inpatient visits Outpatient visits Unadjusted RV1 vaccination cohort Complete vaccination Unvaccinated Adjusteda RV1 vaccination cohort Complete vaccination Unvaccinated Gender Female Male Age (years) 0–1 1–2 2–3 3–4 Year 2008 2009 2010 2011
Adjusteda RV1 vaccination cohort Incomplete vaccination Unvaccinated Gender Female Male Age (years) 0–1 1–2 2–3 3–4 Year 2008 2009 2010 2011
p-value
ER visits
0.43 (0.34, 0.54) Reference
<0.001
1.09 (1.06, 1.12) Reference
<0.001
0.70 (0.65, 0.76) Reference
<0.001
0.37 (0.29, 0.47) Reference
<0.001
1.02 (0.98, 1.06) Reference
0.299
0.62 (0.57, 0.68) Reference
<0.001
0.82 (0.73, 0.92) Reference
0.001
0.86 (0.84, 0.88) Reference
<0.001
0.84 (0.80, 0.89) Reference
<0.001
1.49 (1.32, 1.69) 1.43 (1.27, 1.61) 0.67 (0.57, 0.79) Reference
<0.001 <0.001 <0.001
1.41 (1.37, 1.45) 1.33 (1.30, 1.37) 0.82 (0.80, 0.85) Reference
<0.001 <0.001 <0.001
1.45 (1.37, 1.53) 1.29 (1.22, 1.36) 0.79 (0.74, 0.85) Reference
<0.001 <0.001 <0.001
0.88 (0.78, 1.00) 0.98 (0.90, 1.07) 0.81 (0.73, 0.88) Reference
0.058 0.633 <0.001
1.01 (0.98, 1.04) 1.00 (0.99, 1.02) 0.96 (0.94, 0.97) Reference
0.440 0.626 <0.001
0.81 (0.76, 0.87) 1.04 (1.00, 1.09) 0.89 (0.86, 0.93) Reference
<0.001 0.081 <0.001
Incomplete vaccination versus unvaccinated Inpatient visits Outpatient visits Unadjusted RV1 vaccination cohort Incomplete vaccination Unvaccinated
IRR (95% CI)
ER visits
0.88 (0.63, 1.21) Reference
0.420
1.08 (1.03, 1.14) Reference
0.002
0.95 (0.83, 1.09) Reference
0.448
0.78 (0.52, 1.16) Reference
0.213
1.01 (0.94, 1.08) Reference
0.815
0.83 (0.71, 0.98) Reference
0.029
0.83 (0.73, 0.93) Reference
0.002
0.86 (0.84, 0.88) Reference
<0.001
0.85 (0.81, 0.90) Reference
<0.001
1.47 (1.30, 1.66) 1.47 (1.31, 1.65) 0.68 (0.58, 0.79) Reference
<0.001 <0.001 <0.001
1.42 (1.38, 1.46) 1.34 (1.30, 1.38) 0.82 (0.79, 0.85) Reference
<0.001 <0.001 <0.001
1.45 (1.37, 1.54) 1.30 (1.23, 1.37) 0.79 (0.73, 0.84) Reference
<0.001 <0.001 <0.001
0.89 (0.78, 1.01) 0.98 (0.89, 1.07) 0.78 (0.71, 0.86) Reference
0.081 0.600 <0.001
1.01 (0.98, 1.03) 1.01 (0.99, 1.03) 0.94 (0.92, 0.96) Reference
0.646 0.447 <0.001
0.81 (0.76, 0.86) 1.04 (1.00, 1.09) 0.88 (0.84, 0.92) Reference
<0.001 0.076 <0.001
Abbreviations: IRR, incidence rate ratio; CI, confidence interval; RV1, Rotarix, GSK; ER, emergency room. a Adjusted for age, calendar year and gender.
younger were significantly more likely than children aged 3 to 4 years to experience RV episodes, RV-related inpatient and ER visits, and diarrhea-related visits of any kind, after adjusting for vaccination status, gender and calendar year. Findings from these multivariate analyses provide further evidence that younger patients are at higher risk for RV-related and diarrhea-related events and thus the associated resource utilization. For acute gastroenteritis hospitalizations among children under five years of age in the pre-vaccination era in the US, approximately 40% occurred in children less than 1 year old, approximately 30% in children 12–23 months and approximately 30% among children 24–59 months [19]. The current study also showed there were significant differences in RV episodes, and RV-related and diarrhea-related events by calendar year after adjusting for vaccination status, gender and age. This is consistent with prior studies which have shown year-to-year variability in acute gastroenteritis and rotavirus [14,21]. The current study found statistically significantly lower rates of RV-related resource utilization and diarrhea-related inpatient and ER visit rates for the complete vaccination versus unvaccinated
cohort, yet no difference for diarrhea-related outpatient visits. Other studies on diarrhea-related resource utilization and RV vaccination have yielded mixed findings regarding outpatient visits as well. A similar analysis to the one described here was conducted with the same data source but using data on patients vaccinated with either RV1 or RV5 [9]. Diarrhea-related outpatient visits were significantly higher among the RV vaccinated versus unvaccinated while the opposite was found for inpatient and ER visits; however, this analysis did not adjust for age and calendar year which may have confounded the associations. The study by Cortes et al. [5], with claims data similar to those in the current study, which examined diarrhea-related resource utilization before and after the introduction of RV5 found marked reductions in diarrhea-related hospitalizations following vaccine introduction but only modest, if any, reductions for emergency department and outpatient visits. In their age- and calendar year-adjusted analysis of vaccinated versus unvaccinated children in the era of vaccine availability, rates of diarrhea-related hospitalization, ER and outpatient visits were significantly reduced but substantially moreso for hospitalization and ER. The authors stated that RV accounts for a smaller proportion of
Please cite this article in press as: Krishnarajah G et al. Public health impact of Rotarix vaccination among commercially insured children in the United States. Vaccine (2017), http://dx.doi.org/10.1016/j.vaccine.2017.06.034
7
G. Krishnarajah et al. / Vaccine xxx (2017) xxx–xxx Table 5 Comparison of mean cost of first RV episode and first diarrhea-related episode between vaccinated and unvaccinated cohorts. Cost per 1000 persons mean ($2012)
Unadjusted difference ($2012) (95% CI)
[A]
[A] – [B]
[B]
p-value
Adjusted differencea ($2012) (95% CI)
p-value
Complete vaccination versus unvaccinated Complete vaccination Unvaccinated First RV episode Total costs Inpatient costs Outpatient costs ER costs
788 378 243 167
First diarrhea episode Total costs 83,203 Inpatient costs 26,348 Outpatient costs 43,140 ER costs 13,714
16,040 11,682 2358 2000
15,252 11,305 2114 ( 1833 (
( 19,492; 11,595) ( 15,046; 8036) 2736; 1600) 2243; 1464)
<0.001 <0.001 <0.001 <0.001
11,511 8767 ( 1381 ( 1364 (
( 12,024; 9855) 9563; 7230) 2052; 972) 1932; 1114)
<0.001 <0.001 <0.001 <0.001
147,033 67,028 54,960 25,045
63,831 40,680 11,820 11,331
( ( ( (
<0.001 <0.001 <0.001 <0.001
46,772 37,246 1848 ( 7679 (
( 66,604; 26,268) ( 61,223; 22,895) 5971; 989) 10,563; 6235)
<0.001 <0.001 0.063 <0.001
81,845; 58,305; 16,262; 13,399;
43,928) 22,838) 5668) 9093)
Incomplete vaccination versus unvaccinated Incomplete vaccination Unvaccinated First RV episode Total costs Inpatient costs Outpatient costs ER costs
2492 1893 313 286
First diarrhea episode Total costs 81,963 Inpatient costs 26,374 Outpatient costs 38,994 ER costs 16,595
16,040 11,682 2358 2000
13,548 9789 ( 2045 ( 1714 (
( 18,796; 7586) 14,554; 4358) 2736; 1387) 2212; 1199)
<0.001 0.006 <0.001 <0.001
9809 7228 1323 1258
( ( ( (
147,033 67,028 54,960 25,045
65,070 40,654 15,967 8450 (
( 91,085; 31,167) ( 63,508; 11,511) ( 22,849; 7081) 12,428; 4511)
<0.001 0.014 0.006 0.004
47,703 36,795 5971 ( 4938 (
11,511; 5413) 9355; 2692) 2145; 979) 1364; 235) ( 79,530; 39,433) ( 47,246; 7679) 8900; 1114) 6684; 1232)
<0.001 <0.001 <0.001 <0.001 <0.001 0.004 <0.001 <0.001
Abbreviations: CI, confidence interval; ER, emergency room. a Adjusted for birth year and gender.
diarrhea events in outpatient and ER settings than hospitals. This may render findings based on all diarrhea-related events less informative in some settings, such as outpatient offices. More recently, a study by Leshem et al. which adjusted for age and calendar year reported that outpatient visits for diarrhea among either RV1 or RV5 vaccinated individuals and unvaccinated individuals during the 2010–2011 RV season were similar [10]. Furthermore, a study by Cortese et al. evaluated protection against gastroenteritis among household members where children were vaccinated against RV [20]. In this age-adjusted analysis, outpatient gastroenteritis encounters were modestly increased, whereas inpatient and emergency department visits for gastroenteritis were either decreased or unchanged for household members with an RV vaccinated child versus for household members with an unvaccinated child [20]. The authors speculated there was lower wild-type RV shedding from a vaccinated versus unvaccinated child, and therefore less severe RV cases among other household members of vaccinated children which might lead to more outpatient visits and not acute care visits. Furthermore, in clinical trials, RV vaccination was very effective against hospitalization but far less with mild infections [21]. It is also plausible that vaccinated and unvaccinated patients may seek care for diarrheal diseases in different settings. The economic analysis of RV burden also demonstrated lower cost burden for both complete and incomplete RV1 vaccination relative to unvaccinated children. In the current study’s adjusted analysis, the difference in mean cost for first RV episode was $11.51 for unvaccinated versus completely vaccinated children. Thus, if each of the 34,928 children in the current study who had been vaccinated had not been vaccinated, the cost of first RV episode would have been $402,056 greater for this cohort. There are some limitations of the current study. First, we examined data from only four post-RV1 vaccine seasons (2008–2011), and there were few children in the vaccinated cohorts in the first season. Delayed uptake of RV1 vaccination resulted in follow-up of the vaccinated cohorts having a higher proportion of younger
children with shorter follow-up time relative to the unvaccinated cohort. We cannot be certain that observed changes in the postRV vaccine years were due solely to vaccine use. Secular trends in incidence of RV and other diarrheal pathogens could affect our findings. In addition, unvaccinated children in this study gained herd immunity from those vaccinated with either RV1 or RV5 as both vaccines were available; thus it is difficult to measure the actual impact of RV1 as the unvaccinated cohort receives benefit of immunity from both RV1 and RV5 vaccinated individuals. ICD9 codes were used to identify diagnoses; these codes may not reflect confirmed clinical diagnoses and lack information to assess illness severity. Moreover, medical services obtained outside a patient’s insurance plan are not captured in a claims database. Specifically, RV vaccination obtained outside one’s health plan could not be ascertained and could have resulted in some individuals being misclassified as unvaccinated; this would have biased the results toward the null. There is also potential for underestimating the resource utilization and costs. Finally, inability to adjust for socioeconomic status is a limitation. Given the effectiveness of RV1 vaccination and RV vaccination in general, efforts should focus on improving RV vaccination rates which lag behind most pediatric vaccination rates and continue to fall below the Healthy People 2020 target of 80% [18].
Disclosure statement GK was employed by the GSK group of companies at the time of the study conduct and during the development of the manuscript and is currently employed by CSL Behring. GK also reports ownership of stock options/restricted shares from the GSK group of companies and CSL Behring. AK, CK, MSD, and PL are employees of Analysis Group which has received funding from the GSK group of companies to conduct this study and has received funding for other drug safety and health outcomes research studies by multiple pharmaceutical companies.
Please cite this article in press as: Krishnarajah G et al. Public health impact of Rotarix vaccination among commercially insured children in the United States. Vaccine (2017), http://dx.doi.org/10.1016/j.vaccine.2017.06.034
8
G. Krishnarajah et al. / Vaccine xxx (2017) xxx–xxx
Role of the funding source GlaxoSmithKline Biologicals SA funded this study and was involved in all stages of study conduct, including study design; collection, analysis, and interpretation of the data; writing the report; and the decision to submit the paper for publication. GlaxoSmithKline Biologicals SA also paid all costs associated with the development and publication of this manuscript. Contributions All authors participated in the design or implementation or analysis, and interpretation of the study; and the development of this manuscript. All authors had full access to the data and gave final approval before submission. All authors have agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. Trademark statement Rotarix is a trademark of the GSK group of companies. RotaTeq is a trademark of Merck & Co., Inc. Acknowledgement The authors thank Ning Wu (freelance on behalf of GSK) for writing support and editorial assistance. The authors would like to thank Business & Decision Life Sciences platform for editorial assistance and manuscript coordination, on behalf of GSK. Marie Cloes coordinated manuscript development and provided editorial support. Heather Santiago (publication manager, GSK) also contributed to manuscript coordination and editorial assistance. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.vaccine.2017.06. 034. References [1] Cortese MM, Parashar UD. Centers for Disease Control and Prevention. Prevention of rotavirus gastroenteritis among infants and children: recommendations of the Advisory Committee on Immunization Practices (ACIP). MMWR Recomm Rep 2009;58:1–25. [2] Parashar UD, Alexander JP, Glass RI. Advisory Committee on Immunization Practices Centers for Disease Control and Prevention. Prevention of rotavirus gastroenteritis among infants and children. Recommendations of the Advisory Committee on Immunization Practices (ACIP). MMWR Recomm Rep 2006;55:1–13.
[3] Highlights of prescribing information for Rotarix. Available at:
Please cite this article in press as: Krishnarajah G et al. Public health impact of Rotarix vaccination among commercially insured children in the United States. Vaccine (2017), http://dx.doi.org/10.1016/j.vaccine.2017.06.034