Vm

CPCa

b

c

d

e

a

ARRAA

KSSVPI

1

bcn

t

aP

Cf

0d

Vaccine 29 (2011) 9488– 9492

Contents lists available at SciVerse ScienceDirect

Vaccine

j ourna l ho me pag e: www.elsev ier .com/ locate /vacc ine

alidity of parental report of influenza vaccination in young children seekingedical care�

edric Browna,∗, Haley Clayton-Boswellb, Sandra S. Chavesa, Mila M. Prill a, Marika K. Iwanea,eter G. Szilagyi c, Kathryn M. Edwardsd, Mary A. Staate, Geoffrey A. Weinbergc, Gerry Fairbrothere,aroline B. Hall c, Yuwei Zhud, Carolyn B. Bridgesa, for the New Vaccine Surveillance Network (NVSN)

National Center for Immunization and Respiratory Diseases, Center for Disease Control and Prevention, Atlanta, GA, United StatesVanderbilt Institute for Clinical and Translational Research (VICTR), Vanderbilt University Medical Center, Nashville, TN, United StatesDepartments of Pediatrics, University of Rochester School of Medicine and Dentistry, Rochester, NY, United StatesDepartment of Pediatrics, Vanderbilt University Medical Center, Nashville, TN, United StatesDepartment of Pediatrics, Cincinnati Children’s Hospital Medical Center, Cincinnati, OH, United States

r t i c l e i n f o

rticle history:eceived 19 July 2011eceived in revised form 5 October 2011ccepted 8 October 2011vailable online 19 October 2011

eywords:ensitivitypecificityalidityarental reportnfluenza vaccination

a b s t r a c t

Background: Despite frequent use of self-reported information to determine pediatric influenza vacci-nation coverage, little data are available on the validity of parental reporting of their child’s influenzavaccination status and on factors affecting its accuracy.Methods: We compared parent reported influenza vaccination of children to documented reports of vac-cination collected from medical records (the criterion standard) among children aged 6–59 months whopresented to selected hospitals, emergency departments, and clinics in three U.S. counties with acuterespiratory illness during three influenza seasons (November through May of 2004–2007). Demographicand epidemiologic data were collected from chart reviews and parental surveys.Results: Among 3072 children aged 6–59 months, 47.5% were reported by the parent to have receivedinfluenza vaccine and 39.5% of children had medical record verification of influenza vaccination. Sen-sitivity and specificity of parental reporting was 92.1% and 82.3%, respectively, when compared to theimmunization record. However, 17.7% of children whose parents reported vaccination had no influenzavaccination documented in their medical records, and this proportion was even higher at 28.6%, among

children with an underlying high-risk medical condition. Greater reporting accuracy was associated withyounger age of child (6–23 months vs. 24–59 months), white non-Hispanic race/ethnicity, having healthinsurance, and having a mother with a college education.Conclusions: Our findings indicate that although parental report of influenza vaccination is fairly reli-able (∼76–96%), over reporting by parents often occurs and immunization record review remains the

termi

preferable method for de

. Background

Influenza is a common childhood illness that results in a high

urden of hospitalizations, emergency department and primaryare visits each year [1–10]. The Advisory Committee on Immu-ization Practices (ACIP) first recommended routine influenza

Abbreviations: NVSN, New Vaccine Surveillance Network; PPV, positive predic-ive value; NPV, negative predictive value; ED, emergency department.� Disclaimer: The findings and conclusions of this report are those of the authorsnd do not necessarily represent the views of the Centers for Disease Control andrevention (CDC).∗ Corresponding author at: Centers for Disease Control and Prevention, 1600lifton Road, Mailstop A-34 Atlanta, GA 30333, United States. Tel.: +1 404 639 8590;

ax: +1 404 639 8665.E-mail address: cbrown1@cdc.gov (C. Brown).

264-410X/$ – see front matter. Published by Elsevier Ltd.oi:10.1016/j.vaccine.2011.10.023

ning vaccination status in children.Published by Elsevier Ltd.

vaccination for children 6–23 months of age in 2004, expandedthe recommendation to include children 2–4 years of age in 2006and as of 2010 recommended annual influenza vaccination for allpersons aged 6 months and older [11]. Influenza vaccination is themost effective strategy for reducing the burden of influenza dis-ease [12]. Annual assessment of influenza vaccination coverage iscritical for public health monitoring of vaccine uptake in variousage and risk groups in the population. Accurate ascertainment ofvaccination status is also important for clinicians to determine ifpatients need vaccination and for researchers to conduct vaccineeffectiveness studies.

However, clinicians, researchers and public health programs

often face the challenge of accurately assessing an individual’sreceipt of influenza vaccine. While immunization registries wouldbe the ideal source for such data, in practice, registries are oftenincomplete and may not be representative of the population [13].

ine 29

AleTrsvodica

2

2

StdiN4idld[ia

iw5wicC

2

nohaomhdaPiatditciasth

C. Brown et al. / Vacc

nother challenge is that previously unimmunized children agedess than 9 years need two doses of influenza vaccine which height-ns the challenge of obtaining accurate vaccination information.herefore, clinicians, researchers, and public health programs oftenely on parental report of influenza vaccination [14–19]. Only twotudies have reported on the accuracy of parent recall of influenzaaccination of their children; however, both were limited to onlyne geographic area and lacked generalizability [18,19]. By usingata gathered from active population based surveillance conducted

n 3 U.S. counties, we assessed the validity of parental report of theirhild’s influenza vaccination status over three influenza seasonsnd investigated potential predictors of reporting accuracy.

. Methods

.1. Study population

Data for this cross-sectional study comes from the New Vaccineurveillance Network (NVSN) which conducted active, prospec-ive surveillance for children presenting in hospitals, emergencyepartments (ED), and outpatient clinics for acute respiratory

nfections or fever in the counties that encompass Rochester, NY,ashville, TN, and Cincinnati, OH, USA [1]. Children were enrolled–7 days each week in the hospitals and on rotating days and shifts

n selected EDs. Children were systematically enrolled on selectedays in 8–10 primary care practices in the first year and in one

arge outpatient clinic in each county in the other two years. Furtheretails of the surveillance program have been described elsewhere20]. Institutional Review Boards at the three study sites, the partic-pating surveillance hospitals, and the Centers for Disease Controlnd Prevention approved the study.

This study includes data on children enrolled during threenfluenza seasons (November through May of 2004–2007) who

ere at least 6 months of age on September 1st and no more than9 months of age on November 1st during the influenza season inhich the child was enrolled. Deviations from protocol resulted

n the exclusion of 491 (15.2%) hospitalized children, 19 (0.6%) EDhildren, and 8 (0.2%) clinic children, all of which were from theincinnati site, during the 2005–06 and 2006–07 influenza seasons.

.2. Data collection

Trained interviewers administered a standardized question-aire to the parent or guardian (referred throughout as “parent”)f the child to obtain the child’s demographic information, medicalistory, and influenza vaccination status. Race/ethnicity was char-cterized as non-Hispanic white, non-Hispanic black, Hispanic orther, based on parent report. The parent was also asked about theother’s age and education and the number of children living in the

ousehold. Parents were asked: “Has your child received a flu shoturing the Fall/Winter of (respective influenza season)?” Possiblenswers were “yes”, “no”, “don’t know”, or “refused to respond”.arents were also asked to specify where their child received theirmmunizations (e.g. the name of the clinic or healthcare provider),nd for the first two study seasons, how many doses they received,he date(s) their child received influenza vaccine and, if only oneose was reported, whether their child received influenza vaccine

n a previous year. Health care providers who were specified byhe parent as vaccination providers were contacted to obtain thehild’s immunization record and to collect the number of doses ofnfluenza vaccine given that season, the most recent dose given in

prior season if only one dose was reported in the current sea-on, and the dates of vaccination. The Cincinnati site also accessedhe Ohio state immunization registry in addition to contacting theealth care providers.

(2011) 9488– 9492 9489

Gender, insurance status, and age were obtained from medi-cal record review. Age was recorded as the child’s age on the dateof the admission for inpatients and the date of visit for outpa-tients. Underlying high-risk medical conditions were consideredpresent if noted in the medical record or if the parent/guardianresponded that a health care provider told them the child had thecondition. High-risk conditions included history of asthma, heartdisease, sickle cell anemia, cystic fibrosis, diabetes mellitus, andneurologic conditions, such as seizures, cerebral palsy, or musculardystrophy analogous to the high-risk conditions listed in the ACIPrecommendations for influenza vaccine. History of asthma includesasthma, reactive airway disease, or recurrent or chronic wheezing[11].

2.3. Data analysis

Parent-reported influenza vaccination status of the child wascompared with vaccination status based on the documentedreport of immunization in the patient’s medical record, the cri-terion standard. For the two seasons when parents were askedabout the number of doses received by the child, we checkedparental responses of two doses in the current season against thenumber of doses recorded in the immunization record for that sea-son.

We calculated sensitivity and specificity of parental report ofany influenza vaccine and an unweighted Kappa coefficient toquantify the level of agreement between measures. Sensitivity wasestimated as the percentage of parents who reported any influenzavaccination among those children with a vaccination documentedin their immunization record. Specificity was estimated as thepercentage of parents who reported no vaccination among thosechildren who were unvaccinated according to their immunizationrecord. The positive predictive value (PPV) was estimated as thepercentage of children who were vaccinated as recorded in theimmunization record among those children whose parent reportedthe child as vaccinated. The negative predictive value (NPV) wasestimated as the percentage of children not vaccinated accordingto medical charts among those whose parent reported that thechild did not receive the vaccine. We calculated 95% confidenceintervals (CI) for each estimate using the binomial distribution. Wealso assessed validity of parental report by gender, age (6–23 vs.24–59 months), race/ethnicity, insurance, education of mother, ageof mother, and presence of one or more conditions that increase therisk of influenza-related complications (i.e. high-risk conditions).A chi-square test for difference in proportions was used to detectvalidity values that differ significantly. Statistical analyses wereperformed using SAS 9.2 (Cary, NC) [21].

3. Results

3.1. Demographics

A total of 3225 children were enrolled during the three studyseasons (after excluding the 518 due to protocol deviations fromCincinnati). Of these, a further 153 (4.7%) were excluded from anal-ysis because the parent did not know or respond to questions abouttheir child’s vaccination status. Among the remaining 3072 chil-dren included in the analyses, 53% were male, most of the childrenwere either non-Hispanic black or non-Hispanic white (41% and37% respectively), and 53% were 24–59 months of age. Altogether,

66% of children included in the analyses were privately insured and4.1% were uninsured. An estimated 35% of children had at least onehigh-risk medical condition (Table 1), 43.3% were enrolled in theclinic, 32.2% in the ED, and 24.5% were inpatients.

9490 C. Brown et al. / Vaccine 29 (2011) 9488– 9492

Table 1Potential predictors of reporting accuracy for the study population (2004–05, 2005–06, and 2006–07 seasons combined).

Demographic factor Test characteristic

n % Sensitivity % (95% CI) Specificity %(95% CI)

Positive predictivevalue % (95% CI)

Negative predictivevalue % (95% CI)

Gender Female 1439 46.8 92.1 (89.9–94.3) 83.0 (80.5–85.5) 77.4 (74.2–80.6) 94.3 (92.7–96.0)Male 1633 53.2 92.1 (90.0–94.1) 81.6 (79.2–84.1) 78.2 (75.3–81.0) 93.5 (91.8–95.2)

Age (months) 6–23 1424 46.7 92.2 (90.3–94.1) 77.1 (73.9–80.3)a 82.2 (79.6–84.7)a 89.7 (87.2–92.2)a

24–59 1628 53.3 91.7 (89.2–94.2) 85.1 (83.0–87.1)a 71.4 (67.8–75.0)a 96.2 (95.0–97.4)a

Race/ethnicity White 1134 36.9 94.6 (92.6–96.5)a 82.3 (79.3–85.3)a 81.5 (78.4–84.7)a,b 94.8 (92.9–96.7)Black 1267 41.2 89.9 (87.0–92.7)a 84.9 (82.4–87.3)b 75.6 (71.9–79.3)a 94.1 (92.5–95.8)Hispanic 438 14.3 92.5 (88.9–96.2) 75.1 (69.6–80.6)a,b 75.9 (70.6–81.3) 92.2 (88.5–96.0)Other 233 7.6 87.5 (80.6–94.4) 79.3 (72.7–85.9) 72.0 (63.5–80.5)b 91.3 (86.3–96.2)

Insurance type Private only 2032 66.3 91.5 (89.5–93.4) 82.8(80.7–84.9) 77.1 (74.4–79.8) 93.9 (92.5–95.3)Public/Both 905 29.5 93.2 (90.8–95.7) 81.1 (77.6–84.5) 80.6 (77.0–84.1)a 93.4 (91.1–95.8)Uninsured 127 4.1 90.9 (81.1–1.0) 81.9 (74.1–89.7) 63.8 (50.1–77.6)a 96.3 (92.1–1.0)

Mother’s education <High school 811 26.6 89.8 (86.5–93.2)a 83.9 (80.6–93.0) 78.0 (73.7–82.2) 92.9 (90.5–95.2)High school 920 30.1 91.8 (88.8–94.8) 83.1(80.1–86.1) 74.0 (69.7–78.4)a 95.1 (93.2–96.9)Some college 700 22.9 91.9 (88.8–95.1) 81.4 (77.7–85.2) 77.3 (72.8–81.7) 93.6 (91.1–96.1)College graduate 621 20.3 94.8 (92.4–97.3)a 79.1 (74.6–83.6) 81.9 (77.9–85.9)a 93.9 (91.0–96.8)

Mother’s age <30 years 1943 63.2 91.2 (89.1–93.2) 82.3(80.1–84.4) 75.8 (73.0–78.7)a 93.9 (92.4–95.3)≥30 years 1129 36.8 93.4 (91.2–95.6) 82.4(79.4–85.3) 80.8 (77.6–84.0)a 94.0 (92.0–96.0)

High risk condition Yes 1078 35.1 92.7 (90.4–95.0) 71.4 (67.7–75.1)a 73.3 (69.8–76.7)a 92.1 (89.6–94.5)No 1994 64.9 91.7 (89.7–93.6) 87.4 (85.5–89.2)a 81.1 (78.5–83.8)a 94.6 (93.3–95.9)

S tor foa panic

3i

oftd4edt

couT2mw9ws

sfcPoattohariaafs

tatistically significant comparisons between superscripted levels of each patient facnd Hispanic children (both denoted by superscript “a”) and between Black and His

.2. Measures of validity of parental report by agreement withmmunization records

The percentage of children vaccinated against influenza basedn medical record verification increased over the three seasons,rom 36% during 2004–05 to 47% during the 2006–07 season. Forhe 3 years combined, 40% of enrolled children received one or moreoses of influenza vaccine according to immunization records while7% of parents reported that their child had been vaccinated. Par-nts recalled a higher proportion of children receiving at least oneose (47.5% vs. 39.5%) and of being fully vaccinated (21.5% vs. 20.0%)han were verified by medical records.

Sensitivity and specificity of parental report of a child’s vac-ination were 92.1% and 82.3% respectively for the three seasonsverall. Thus, 17.7% of children with no influenza vaccination doc-mented in their medical records had parent reported vaccination.his proportion was higher among the 1078 high risk children;8.6% of high risk children with no influenza vaccination in theiredical records had parent reported vaccination. The overall PPVas 77.8% (range 75.9–80.3% across seasons). The overall NPV was

3.9% (range 91.2–95.8% across seasons). An overall kappa scoreas 72.2% and ranged between 69.2 and 73.1% across the three

easons (Table 2).Table 1 shows results of reporting accuracy by demographic

trata for the three seasons combined (similar results were notedor individual seasons, data not shown). There were no signifi-ant observed differences associated with a child’s gender. ThePV was statistically significantly higher for children 6–23 monthsf age vs. 24–59 months (82.2% vs. 71.4% respectively). Childrenged 24–59 months had higher validity for reports of not receivinghe influenza vaccine for both specificity (85.1% vs. 77.1%, respec-ively) and NPV (96.2% vs. 89.7%, respectively). Parental reportingf child’s vaccination was more accurate if the child did not have aigh-risk condition (PPV 81.1% vs. 73.3%; p < 0.001). Higher PPV waslso observed among non-Hispanic whites as compared to otheraces/ethnicities, among children who were privately or publiclynsured verses uninsured, and for children whose mother gradu-

ted from college. Sensitivity of parental reporting declined to 58%nd PPV to 48.2% as compared to medical records when we assessedull vaccination, i.e., children receiving 2 doses during the currenteason or receiving one dose in current season but with a history

r given test characteristic; e.g., specificity was significantly different between White children (both denoted by superscript “b”).

of a dose in the previous season. However, among these childrenthe specificity and NPV of parental reporting remained relativelyhigh at 82.1% and 87.2%, respectively, indicating that parents couldrecall reasonably well if their children were not up-to-date withinfluenza vaccinations.

4. Discussion

In order to improve convenient access to influenza vaccine andto boost vaccination rates, influenza vaccines are becoming moreavailable outside medical care settings, such as in schools andretail pharmacies. Thus, vaccine providers may have to increas-ingly rely on parental report of a child’s vaccination status. Inour study, parental reporting of no influenza vaccination was anaccurate predictor of not being vaccinated (NPV 93.9%). However,17.7% of parents reported their child as being vaccinated againstinfluenza when the immunization record indicated no vaccination.Understanding the likely accuracy of parental report of influenzavaccination is important to guide medical decisions about the needfor vaccination and also for interpreting influenza vaccine coveragedata that relies on parent report.

Our findings were comparable to those reported by Shinall et al.from a smaller study of 198 children seen in a single primary careclinic population in Tennessee during influenza season in 2005[18]. However, our study differed from a study by Irving et al.which included parent-reported vs. registry-based influenza vac-cination status of children aged 6–59 months enrolled with acuterespiratory illness. Their study reported a higher specificity (92%)and positive predictive value (94%) than estimated in our study(82% and 78%, respectively) [19]. However, the study populationwas predominantly rural, mostly white, and received inpatient andoutpatient care from a single healthcare organization in Wiscon-sin – thus, it may be difficult to generalize their findings to otherpopulations.

In our study, parents of children with high risk conditions, forwhom vaccine is especially important due to their increased risk ofsevere influenza illness, were less likely to report influenza vac-

cination accurately than parents of healthy children (specificityof 71% vs. 87%, respectively). This could be because children withconditions like diabetes, asthma, or heart disease, have more med-ical encounters and more healthcare providers than their healthy

C. Brown et al. / Vaccine 29 (2011) 9488– 9492 9491

Table 2Comparison of parent report of influenza vaccination with immunization record, by season for children aged 6–59 months.

Parent report Medical record Test characteristic

Yes No Total Sensitivity %(95% CI)

Specificity %(95% CI)

Positivepredictivevalue % (95% CI)

Negativepredictivevalue % (95% CI)

Kappa % (95% CI)

Over all Yes 1139 325 1464 92.1 (90.6–93.6) 82.3 (80.5–84.0) 77.8 (75.7–79.9) 93.9 (92.7–95.1) 72.2 (69.8–74.6)No 98 1510 1608Total 1237 1835 3072

2004–05 Yes 528 168 696 93.5 (91.4–95.5) 83.4 (81.1–85.7) 75.9 (72.7–79.0) 95.8 (94.5–97.1) 73.1 (69.7–76.5)No 37 842 879Total 565 1010 1575

2005–06 Yes 277 68 345 90.2 (86.9–93.6) 83.6 (80.0–87.1) 80.3 (76.1–84.5) 92.0 (89.3–94.8) 72.6 (67.6–77.7)No 30 346 376Total 307 414 721

2006–07 Yes 334 89 423 91.5 (86.6–94.4) 78.4 (74.4–82.3) 79.0 (75.1–82.8) 91.2 (88.3–94.2) 69.2 (64.2–74.2)No 31 322 353Total 365 411 776

N

crdc9ttcsasb

cven6(3bccgff

otihardvavocpdmAmhu

ote: there were no statistically significant differences between years.

ounterparts and their parents, therefore, may be more prone toecall biases. Shinall et al. reported a specificity of 81% for both chil-ren aged 6–23 months and children aged 24–59 months havinghronic medical condition. However, this was a decline from the7% specificity of healthy children aged 24–59 months reported inhe same study. One other study that assessed parental perspec-ives on influenza vaccination of children with chronic medicalonditions found even lower specificity (68.7%) than ours. Thistudy, however, collected data on older children (up to 13 years ofge) from low-income urban neighborhood clinics and used a mailurvey, factors that may in part explain the differences in resultsetween the studies [22].

Based on medical record review, we found that only 38–44% ofhildren who were age-eligible for vaccination received influenzaaccine in our study population over the 3 year study period. Ourstimates for receipt of at least one vaccination were higher thanational estimates, which were 33% (95% CI; 32–35%) for children–23 months of age from the 2005 National Immunization SurveyNIS), 32% (95% CI; 31–34%) from the 2006 NIS, and 32% (95% CI;0–33%) from the 2007 NIS [23–25]. The higher estimates coulde due to differences in populations, with our study consisting ofhildren seeking medical care, possibly with a higher proportion ofhildren having pre-existing high risk conditions compared to theeneral population. There was no significant difference in estimatesor fully vaccinated children when comparing the NIS and our studyor years 2004–2006.

In our study population, the child’s age, race/ethnicity, typef insurance, high risk condition, and mother’s age and educa-ion were factors in the validity of parental reporting for influenzammunizations. However, specificity and NPV remained relativelyigh consistently across seasons and demographic factors. Thus,mong children seen in any medical setting where the parenteports no vaccination, providers may have relatively high confi-ence of no prior influenza vaccination and administer influenzaaccine, especially given the strong safety profile of both livend inactivated influenza vaccines for approved age groups andery low risk of adverse events from receipt of an extra dosef vaccine [11]. When counseling parents about influenza vac-ination, providers who are concerned about the accuracy of aarental report of vaccination should, especially for high risk chil-ren, take into account that a substantial number of these childrenay not in fact be vaccinated or not be fully vaccinated [26,27].

s local and state health departments consider the benefits ofore non-traditional vaccine delivery settings for improving child-

ood immunization rates, including schools and pharmacies, these of immunization registries integrating non-traditional settings

becomes more valuable [27] and can benefit patient care at theprovider level.

A strength of our study was the inclusion of three influenzaseasons as vaccination recommendations were expanding, sothat measures of validity and agreement could be compared byinfluenza season. Despite changes in vaccine recommendations,the accuracy of parental report was similar over the three studyyears which may indicate that these findings might be extrapolatedacross seasons. Evaluation of additional years after implementa-tion of the universal influenza vaccination recommendation wouldbe helpful to assess trends over time on reporting accuracy andconfirm our finding. In contrast to other studies’ findings, ours arepossibly more generalizable due to the diversity of the study pop-ulation drawn from three different geographic regions. Limitationsof this study include the potential for recall bias since enrolledchildren are more likely to seek medical care because of their ill-ness; perhaps parents of children with access to healthcare havebetter recollection of their vaccination history which could makeour results an overestimation of parent reporting accuracy. Addi-tionally, although study personnel attempted to verify vaccinationstatus for all children, some vaccinations may have been missed.While the medical record is considered the gold standard, there isalways a possibility that lack of documentation of vaccination in therecord reviewed does not accurately reflect vaccination status dueto multiple providers or vaccine uptake outside of the medical caresetting. Another limitation is the exclusion of 518 children from thestudy due to a protocol deviation at the Cincinnati site during the2005–06 and 2006–07 seasons. Although there is a potential for biasdue to excluding these children, those excluded did not differ, usingthe .05 significance level, from those included with respect to age,sex, race, and insurance status. In addition, all analyses were runwith and without the Cincinnati data and there were no significantdifferences in the validity estimates or their confidence intervals.

5. Conclusions

Although a parent’s report of the receipt of any dose of influenzavaccine given to their child in the same year closely agreed withdocumentation in the child’s immunization record (92.1% sensi-tivity), the accuracy in recalling past immunizations and assessinga need for second dose was less sensitive. Therefore, verification

of vaccination history by immunization record review remains thepreferable method for determining vaccination status of a child,particularly when considering the need for two vaccine doses foryoung children. Immunization registries when fully implemented

9 cine 29

hnan

A

sGUdJC

iddEBoBclaiFIhcdWaRhtUmUsa

R

[

[

[

[

[

[

[

[

[

[

[

[[

[

[

[

[

492 C. Brown et al. / Vac

ave the potential to provide timely and accurate data if vacci-ation providers fully participate. As more influenza vaccinationsre given in non-traditional settings, these additional facilities willeed to be integrated into the vaccination reporting systems.

cknowledgements

The authors acknowledge the contribution of the following per-ons to surveillance-related activities: University of Rochester:eraldine Lofthus, Kenneth Schnabel, Andrea Marino; Vanderbiltniversity: Diane Kent, Yi Wei Tang, John Williams; Cincinnati Chil-ren’s Medical Center: Diana Henderson, Michol Holloway, Linda

amison, David Witte, Joel Mortensen, Monica McNeal; CDC: Aaronurns, Minnie Wang, Ranee Seither, Jennifer Reuer.

Contributions: Mr. Brown had full access to all of the datan the study and takes responsibility for the integrity of theata and the accuracy of the data analysis. Study concept andesign: Brown, Clayton-Boswell, Bridges, Iwane. Acquisition of data:dwards, Szilagyi, Staat, Zhu. Analysis and interpretation of data:rown, Clayton-Boswell, Bridges, Chaves, Prill, Iwane. Draftingf the manuscript: Clayton-Boswell, Brown, Chaves, Iwane, Prill,ridges. Critical revision of the manuscript for important intellectualontent: Brown, Clayton-Boswell, Chaves, Iwane, Bridges, Prill, Szi-agyi, Edwards, Staat, Weinberg, Fairbrother, Hall, Zhu. Statisticalnalysis: Brown. Obtained funding: Edwards, Szilagyi, Staat. Admin-strative, technical, or material support: Edwards, Szilagyi, Staat,airbrother, Iwane, Prill, Zhu. Study supervision: Brown, Chaves,wane, Bridges. Conflicts of interest: The authors declare that theyave no conflicts of interest except for the follow financial dis-losures. Financial disclosures: The following authors have madeisclosure: K. Edwards received grant funding from Novartis. G.einberg was on speaker’s bureaus of Merck, GlaxoSmithKline

nd Sanofi Pasteur. M. Staat had funding from MedImmune forSV studies and was on the MedImmune Advisory Board. C. Hallas been on a MedImmune Advisory Board and is a consultanto MedImmune. Funding/support: The work was supported by theS Centers for Disease Control and Prevention [cooperative agree-ent numbers U38/CCU217969, U01/IP000017, U38/CCU417958,01/IP000022, U38/CCU522352, U01/IP000147]. Role of the spon-

or: CDC provided funding support. The study had CDC co-authorsnd CDC staff reviewed the manuscript. Clinical centers: None.

eferences

[1] Poehling KA, Edwards KM, Weinberg GA, Szilagyi P, Staat MA, Iwane MK, et al.The underrecognized burden of influenza in young children. N Engl J Med2006;355(1):31–40.

[2] Izurieta HS, Thompson WW, Kramarz P, Mitchel Jr EF, Griffin MR. Influenza and

the rates of hospitalization for respiratory disease among infants and youngchildren. N Engl J Med 2000;342:232–9.

[3] Neuzil KM, Mellen BG, Wright PF, Mitchel Jr EF, Griffin MR. The effect ofinfluenza on hospitalizations, outpatient visits, and courses of antibiotics inchildren. N Engl J Med 2000;342:225–31.

[

(2011) 9488– 9492

[4] Glezen WP, Neuzil KM, Griffin MR, Wright PF, Izurieta HS, Thompson WW, et al.Influenza and hospitalizations in Children. N Engl J Med 2000;342:1752–3.

[5] Neuzil KM, Zhu Y, Griffin MR, Edwards KM, Thompson JM, Tollefson SJ, et al.Burden of interpandemic influenza in children younger than 5 years: a 25-yearprospective study. J Infect Dis 2002;185:147–52.

[6] Grijalva CG, Weinberg GA, Bennett NM, Staat MA, Craig AS, DuPont WD, et al.Estimating the undetected burden of influenza hospitalizations in children.Epidemiol Infect 2007;135(6):951–8.

[7] Watkins J. Influenza: burden of disease in childhood. Int Congr Ser2004;1263:263–6.

[8] Mullooly JP, Barker WH. Impact of type A influenza on children: a retrospectivestudy. Am J Public Health 1982;72(9):1008–16.

[9] Couch RB. Influenza: prospects for control. Ann Intern Med2000;133(12):992–8.

10] Miller EK, Griffin MR, Edwards KM, Weinberg GA, Szilagyi PG, Staat MA, et al.New Vaccine Surveillance Network. Influenza burden for children with asthma.Pediatrics 2008;121(1):1–8.

11] Centers for Disease Control and Prevention. Prevention and control of sea-sonal influenza with vaccines: recommendation of the Advisory Committeeon Immunization Practices (ACIP), 2010. MMWR 2010;59(No. RR-08):1–62.

12] National Advisory Committee on Immunization. Statement on influenza vac-cination for the 2008–2009 season. An Advisory Committee Statement (ACS).Can Commun Dis Rep 2008;34:1–46.

13] Placzek H, Madoff LC. The use of immunization registry-based data in vaccineeffectiveness studies. Vaccine 2011;29(3):399–411.

14] Lu PJ, Ding H, Euler GL, Furlow C, Bryan LN, Bardenheier B, et al. Interim results:state-specific influenza A (H1N1) 2009 monovalent vaccination coverage –United States, October 2009–January 2010. MMWR 2010;59(12):363–8.

15] Zimmerman LA, Bartlett DL, Enger KS, Gosney K, Williams WG. Influenza vac-cination coverage: findings from immunization information systems. BMCPediatrics 2007;7:28.

16] Gargiullo P, Shay D, Katz J, Bramley A, Nowell M, Michalove J, et al. Effec-tiveness of 2008–09 trivalent influenza vaccine against 2009 pandemicinfluenza A (H1N1) – United States, May–June 2009. Morb Mortal Wkly Rep2009;58(44):1241–5.

17] Bolton P, Holt E, Ross A, Hughart N, Guyer B. Estimating vaccination coverageusing parental recall, vaccination cards, and medical records. Public Health Rep1998;113(6):521–6.

18] Shinall Jr MC, Plosa EJ, Poehling KA. Validity of parental report of influenzavaccination in children 6 to 59 months of age. Pediatrics 2007;120(4):783–7.

19] Irving SA, Donahue JG, Shay DK, Ellis-Coyle TL, Belongia EA. Evaluation of self-reported and registry-based influenza vaccination status in a Wisconsin cohort.Vaccine 2009;27:6546–9.

20] Eisenberg KW, Szilagyi PG, Fairbrother G, Griffin MR, Staat M, Shone LP, et al.Vaccine effectiveness against laboratory-confirmed influenza in children 6 to59 months of age during the 2003–2004 and 2004–2005 influenza seasons.Pediatrics 2008;122:911–9.

21] SAS Institute Inc. Cary, NC, USA. Version 9.2, 2002–2008.22] Lin CJ, Zimmerman RK, Nowalk MP, Ko FS, Raymund M, Hoberman A, et al.

Parental perspectives on influenza vaccination of children with chronic medicalconditions. J Natl Med Assoc 2006;98(2):148–53.

23] Centers for Disease Control and Prevention. Childhood influenza vac-cination coverage – United States, 2004–05 influenza season. MMWR2006;59(39):1062–5.

24] Centers for Disease Control and Prevention. Influenza vaccination coverageamong children aged 6–23 months – United States, 2005–06 influenza season.MMWR 2007;56(37):959–63.

25] Centers for Disease Control and Prevention. Influenza vaccination coverageamong children aged 6–23 months – United States, 2006–07 influenza season.MMWR 2008;57(38):1039–43.

26] Committee on practice and ambulatory medicine. Immunization information

systems. Pediatrics 2006;118:1293.

27] Pickering LK, Baker CJ, Freed GL, Gall SA, Grogg SE, Poland GA, et al. Immuniza-tion programs for infants, children. adolescents, and adults: clinical practiceguidelines by the Infectious Disease Society of America. Clin Infect Dis2009;49(6):817–40.