Volume 12, Number 10—October 2006
Research
Health Benefits, Risks, and Cost-Effectiveness of Influenza Vaccination of Children
Table A1
Variable | Most likely estimate | Range for sensitivity analysis | Source | Type of distribution | Distribution parameter 1 | Distribution parameter 2 | |
---|---|---|---|---|---|---|---|
Influenza illness attack rate (annual) | (1–10) | ||||||
6–23 mo | 0.157 | 0.02–0.35 | β1 | 2.2 | 11.8 | ||
2 y | 0.155 | 0.02–0.35 | Derived2 | ||||
3–4 y | 0.155 | 0.02–0.35 | Derived | ||||
5–11 y | 0.08 | 0.01–0.18 | Derived | ||||
12–17 y | 0.06 | 0.01–0.14 | Derived | ||||
Probability of an outpatient visit for child with influenza illness3 | (5,11,12)4 | ||||||
6–23 mo | 0.5 | 0.17–0.83 | β | 3.3 | 3.5 | ||
2 y | 0.47 | 0.15–0.81 | β | 3.29 | 3.71 | ||
3–4 y | 0.43 | 0.12–0.78 | β | 3.01 | 3.99 | ||
5–11 y | 0.28 | 0.11–0.5 | β | 5.6 | 14.4 | ||
12–17 y | 0.24 | 0.06–0.5 | β | 2.88 | 19.12 | ||
Probability of otitis media for a child with medically attended influenza illness | (13–16), expert panel | ||||||
6–23 mo | 0.63 | 0.33–0.8 | β | 6.3 | 3.7 | ||
2 y | 0.58 | 0.27–0.8 | β | 5.22 | 3.78 | ||
3–4 y | 0.39 | 0.17–0.6 | β | 6.24 | 9.76 | ||
5–11 y | 0.23 | 0.05–0.5 | β | 2.53 | 8.47 | ||
12–17 y | 0.15 | 0.01–0.4 | β | 1.5 | 8.5 | ||
Probability of nonhospitalized pneumonia or other outpatient complication for child with medically attended influenza illness5 | (11,12); expert panel | ||||||
6–23 mo | 0.2 | 0.04–0.5 | β | 2.6 | 10.4 | ||
2 y | 0.15 | 0.02–0.4 | β | 1.95 | 11.05 | ||
3–4 y | 0.15 | 0.02–0.4 | β | 1.95 | 11.05 | ||
5–11 y | 0.11 | 0.02–0.3 | β | 2.2 | 17.8 | ||
12–17 y | 0.08 | 0.01–0.2 | β | 2.16 | 24.84 | ||
Hospitalizations for pneumonia or other respiratory conditions due to influenza per 10,000 children not at high risk6 | (7,11,17); W. Thompson, pers. comm.) | ||||||
6–23 mo | 28.3 | 1.9–80.0 | β | 5.5 | 244.5 | ||
2 y | 17.1 | 0–56.8 | β | 3.4 | 246.6 | ||
3–4 y | 8.0 | 0–35.4 | β | 1.6 | 248.4 | ||
5–11 y | 3.1 | 0–16.0 | β | 7.95 | 1,492.1 | ||
12–17 y | 3.1 | 0–14.9 | β | 10.5 | 1,489.5 | ||
Probability of long-term sequelae after influenza-related hospitalization2 | 0.01 | 0.001–0.03 | Expert panel | β | 1.3 | 11.7 | |
Probability of death during influenza-related hospitalization | 0.0009 | 0–0.002 | (18)4 | β | 1.7 | 18.3 | |
Vaccine effectiveness in preventing influenza illness9 | |||||||
IIV | 0.69 | 0.4–0.9 | (19)4 | β | 7.59 | 3.41 | |
LAIV | 0.838 | 0.6–0.96 | (20)4 | β | 16.76 | 3.24 | |
Probability of medically attended vaccination-related adverse events | |||||||
Injection site reaction | |||||||
6–23 mo | 0.008 | 0.002–0.017 | (8) | β | 4.0 | 46.0 | |
2 y | 0.003 | Derived10 | |||||
3–4 y | 0.002 | Derived | |||||
5–11 y | 0.001 | Derived | |||||
12–17 y | 0.0003 | Derived | |||||
Systemic reaction (fever)11 | |||||||
6–23 mo | 0.013 | 0.001–0.025 | (20) | β | 5.2 | 194.8 | |
2 y | 0.011 | Derived | |||||
3–4 y | 0.009 | Derived | |||||
5–11 y | 0.004 | Derived | |||||
12–17 y | 0.003 | Derived | |||||
Anaphylaxis | 0.00000025 | 0–0.000001 | Expert panel | β12 | 0.5 | 19.5 | |
Guillain-Barré syndrome | 0.000001 | 0–0.00001 | Expert panel | Triangular | 0.000001 (most likely) | 0 (min), 0.000002 (max) | |
Influenza-related costs | |||||||
OTC medications13 | $3 | (21,22); J. Finkelstein, pers. comm.; expert panel | |||||
Physician visit for uncomplicated influenza14 | $27 | $0–$180 | Marketscan database15 | Lognormal16 | 32 | 27 | |
Physician visit for otitis media | |||||||
6–3 mo | $78 | $23–$197 | Marketscan database17 | Lognormal | 98 | 78 | |
2–4 y | $83 | $23–$200 | Marketscan database17 | Lognormal | 100 | 83 | |
5–17 y | $94 | $31–$245 | Marketscan database17 | Lognormal | 117 | 94 | |
Physician visit for nonhospitalized pneumonia | |||||||
6–23 mo | $179 | $62–$715 | Marketscan database17 | Lognormal | 252 | 179 | |
2–4 y | $88 | $27–$333 | Marketscan database17 | Lognormal | 130 | 88 | |
5–17 y | $109 | $34–$503 | Marketscan database17 | Lognormal | 187 | 109 | |
Hospitalization18 | |||||||
6–23 mo | $4,306 | $1,307–$34,473 | Marketscan database17 | Lognormal | 13194 | 4306 | |
3–4 y | $4,180 | $1,292–$32,030 | Marketscan database17 | Lognormal | 10000 | 4180 | |
5–17 y | $5,135 | $1,373–$42,990 | Marketscan database17 | Lognormal | 14956 | 5135 | |
Long-term sequelae following influenza-related hospitalization19 | $625,000 | $0–$1,000,000 | (23) | ||||
Vaccination costs | |||||||
Per dose, IIV20 (children <3 y) | $9.5621 | 1×–4× base case | (21) | ||||
Per dose, IIV (children >3 y) | $6.8621 | 1×–4× base case | (21) | ||||
Per dose, LAIV20 | $12.8922 | $10–$25 | (24,25) | ||||
Administration (0–2 visits)23 | $25 | $10–$40 | (26) | ||||
Parent time costs24 | $32 | $0–$62 | (27), expert panel | ||||
Total vaccination costs | $30–$110 | ||||||
6–23 mo | $79 | ||||||
2 y | $66 | ||||||
3–4 y | $59 | ||||||
5–11 y | $49 | ||||||
12–17 y | $49 | ||||||
Vaccination-related adverse events | |||||||
Physician visit for injection site reaction25 | $61 | $30–$-683 | Marketscan database26 | Lognormal16 | 202 | 61 | |
Anaphylaxis27 | $2,699 | $52–$13,754 | Marketscan database28 | Lognormal16 | 4527 | 2699 | |
Guillain-Barré syndrome29 | $23,359 | $6,663–$78,912 | Marketscan database28 | Lognormal16 | 32196 | 23359 | |
Quality adjustments30,31 (disutility associated with an event) | |||||||
Episode of influenza | 0.005 | 0.002–0.009 | (27) | β | 7.35 | 1492.65 | |
Otitis media | 0.042 | 0.023–0.065 | (28) | β | 14.56 | 335.44 | |
Nonhospitalized complications (pneumonia) | 0.046 | 0.027–0.071 | (28) | β | 16.21 | 333.8 | |
Hospitalization, pneumonia | 0.076 | 0.054–0.100 | (28) | β | 37.85 | 462.15 | |
Anaphylaxis | 0.02 | 0.006–0.041 | (27) | β | 4.53 | 225.47 | |
Guillain-Barré syndrome | 0.141 | 0.092–0.199 | (27) | β | 22.53 | 137.47
IIV, inactivated influenza vaccine; LAIV, live, attenuated influenza vaccine; OTC, over the counter. |
IIV, inactivated influenza vaccine; LAIV, live, attenuated influenza vaccine; OTC, over the counter.
1Distributions for transition probabilities were assigned using most likely values and ranges identified in the literature and/or expert panel. For these parameters, primary data were not available and beta distributions were assigned to match the values identified in the table.
2Distributions for age groups other than 6–23 mo are based on the 6- to 23-mo distribution multiplied by the ratio of the most likely estimates for the age group in question to children 6–23 mo (e.g., the distribution for 2 y is calculated by multiplying the distribution for 6–23 mo by 0.155/0.157).
3Estimates for healthy children are shown in Table. Probabilities are estimated to be twice as high for children at high risk for influenza-related complications.
4Range for sensitivity analysis determined by expert opinion.
5Estimates for healthy children shown in Table. Probabilities are estimated to be up to 5 times as high for children at high risk for influenza-related complications. Base case estimates for children at high risk are 1.6 times as high as for healthy children.
6Children at high risk are estimated to be hospitalized at 3–6 times the rate of healthy children.
7Probability from distribution divided by 10.
8Probability from distribution divided by 100.
9Assumes vaccine is poorly matched with circulating virus 1 in 10 years (i.e., vaccine effectiveness is assumed to be 0 in years with a poor match).
10Distributions for age groups other than 6–23 mo are based on the 6- to 23-mo distribution multiplied by the ratio of the most likely estimates for the age group in question to children ages 6–23 mo (e.g., the distribution for 2 years is calculated by multiplying the distribution for 6–23 mo by 0.003/0.008).
11Definitions and follow-up for incidence of fever following vaccination vary by study. Rates are 2× higher for high-risk subgroups.
12Probability from distribution divided by 100,000.
13Vary by age, calculated by costing out recommended dose of acetaminophen for average weight in each age group.
14Only a proportion of children with influenza illness are assumed to make a physician visit. ICD-9 codes: 487 and 487.0.
151993–1997 Marketscan database, The Medstat Group, Ann Arbor, MI, USA.
16Lognormal distributions are approximated using the mean and median in Treeage. In this table, parameter 1 is the mean and parameter 2 is the median for each distribution.
172001-2003 Marketscan database, The Medstat Group, Ann Arbor, MI.
18ICD-9 codes: 460-466, 471-474, 477, 478, 480-483, 490-496, 506-508, 510, 511, 514, 518, 519. 2001-2003 Marketscan database.
19Includes costs of lifetime care and special education.
20Assumed 2 doses will be required for children <5 years receiving their first influenza vaccination.
21Vaccine dose costs are based on 2004 CDC negotiated prices. Cost for children <3 years assumes thimerosal-free vaccine is used.
22Based on 2004 CDC negotiated price.
23Common Procedural Terminology (CPT) codes: 99211, 90471. Physician costs for vaccine administration at existing visit is $10.37 (90471); $19.95 for vaccine administration requiring a separate visit (99211).
24Each physician visit is assumed to take 2 hours of parent time valued at an average hourly wage rate of $15.54.
255- minute visit, CPT code 99211.
26 2001–2003 Marketscan database.
27ICD-9 codes: 999.4, 995.0, 995.6x.
282001-2003 Marketscan database.
29ICD-9 code: 357.0.
30Quality adjustments are included in the model as a one-time decrement in utility for each temporary health state. For example, an episode of influenza results in a one-time loss of 0.005 quality-adjusted life years (QALYs). Utility losses were calculated by dividing the discounted time-traded off by the respondent’s discounted life expectancy.
31Average life span used to calculate total QALYs lost due to life-long sequelae and death was 77.9–78.2 y, depending on child’s current age. See Table A1 References in Appendix.