National Impact of Universal Childhood Immunization With Pneumococcal Conjugate Vaccine on Outpatient Medical Care Visits in the United States
BACKGROUND. Since introduction of the heptavalent pneumococcal conjugate vaccine in the United States in 2000, rates of invasive pneumococcal disease have declined. However, the national impact of heptavalent pneumococcal conjugate vaccine on pneumonia and otitis media remains unknown.
OBJECTIVES. We compared national rates of outpatient visits for pneumonia and otitis media in children before and after heptavalent pneumococcal conjugate vaccine introduction.
METHODS. Rates of ambulatory visits for pneumococcal and nonspecific pneumonia, otitis media, and other acute respiratory infections were compared before (1994–1999) and after (2002–2003) heptavalent pneumococcal conjugate vaccine introduction using the National Ambulatory Medical Care Survey and the National Hospital Ambulatory Medical Care Survey. To evaluate vaccine effects while accounting for temporal variability, ratios of pneumococcal-related disease rates in children <2 years old (vaccine target population) and in children 3 to 6 years old (not routinely vaccinated) were evaluated using a Poisson regression analysis. For children <2 years old, the differences between observed and expected rates were the estimated vaccine effects.
RESULTS. After the introduction of heptavalent pneumococcal conjugate vaccine, otitis media visit rates declined by 20% in children aged <2 years. This decline represented 246 fewer otitis media visits per 1000 children aged <2 years annually. There were no significant decreases in outpatient visit rates for pneumonia or other acute respiratory infections for children aged <2 years.
CONCLUSIONS. After heptavalent pneumococcal conjugate vaccine introduction, national rates of otitis media visits declined significantly in children <2 years old. Persistence of this trend will produce a significant reduction of the otitis media burden and further enhance the cost-effectiveness of heptavalent pneumococcal conjugate vaccine.
Streptococcus pneumoniae is an important cause of pneumonia and otitis media in children in the United States.1–6 Although the precise incidence of pneumococcal pneumonia is difficult to ascertain, because routine diagnostic tests are insufficiently specific and sensitive,1S pneumoniae likely accounts for 24% to 33% of pneumonias in ambulatory settings.6–8 Children <5 years of age have the highest incidence of pneumonia,4,5,9 and pneumonia and influenza combined represent the leading infectious cause of mortality in children aged 1 to 4 years in the United States.10 Worldwide, ∼1.9 million annual childhood deaths are attributed to pneumonia.11,12
Otitis media, the most common cause of sick visits in preschool-aged children in the United States, accounts for an estimated 24.5 million clinic visits and $2 billion to $5.3 billion in annual treatment costs.13–20 By the age of 3 years, 83% of children will have experienced ≥1 episode of otitis media. Because otitis media is the most common reason for antibiotic prescriptions among US children,19–21 it is an important contributor to antibiotic resistance.22S pneumoniae accounts for 30% to 55% of otitis media episodes.3,18,23
In 2000, a heptavalent pneumococcal conjugate vaccine (PCV7) was licensed in the United States and recommended for use in all children <2 years of age.1,24 Results from randomized, controlled clinical trials showed significant reductions in rates of invasive disease, pneumonia, and otitis media in fully vaccinated children.11,16,23,25–29 Post-PCV7 licensure studies have shown significant decreases in the rates of invasive pneumococcal disease, with the largest reductions in rates observed in children <2 years of age.30–32 These reductions have occurred despite vaccine shortages and before full implementation of recommendations for universal vaccination.
Because the incidence of pneumococcal pneumonia and otitis media is much higher than invasive disease, PCV7 might have a more profound impact on the epidemiology and health care costs associated with these noninvasive pneumococcal-related diseases than those seen with invasive diseases. Furthermore, cost-effectiveness evaluations of immunization programs are more sensitive to estimates of PCV7 effects on pneumonia and otitis media rates than those of invasive diseases.33,34 However, initial cost-effectiveness evaluations of PCV7 did not consider vaccine indirect effects,33,35 and nationally representative data on the impact of PCV7 in decreasing pneumonia and otitis media are limited.16,18,31,33
Six years after the introduction of PCV7, we evaluated the potential impact of routine PCV7 immunization on visit rates for 2 common pneumococcal-related diseases, pneumonia and otitis media, in the US outpatient setting through the analysis of 2 National Health Care Surveys.
We analyzed data collected by the National Ambulatory Medical Care Survey (NAMCS) and the National Hospital Ambulatory Medical Care Survey (NHAMCS) from 1994 to 2003. These surveys provide standardized and detailed data on outpatient diagnoses and treatments annually without any personal identifiers.36–39 Up to 3 medical diagnoses per visit are coded in NAMCS and NHAMCS, using the International Classification of Diseases, Clinical Modification, Ninth Revision (ICD-9-CM). To account for this multiplicity, 3 mutually exclusive diagnosis groups were created using a prespecified hierarchy: pneumococcal pneumonia and nonspecific pneumonias, otitis media, and other acute respiratory infections (ARIs) (see Appendix for specific diagnosis codes). Some sampled visits represented new episodes and others follow-up visits. For the years 2001–2003, which included visit encounter type information, the proportion of pneumonia visits that were classified as follow-up visits ranged from 24% to 59% for children aged <2 years. Similarly, the proportion of otitis media and other respiratory infection visits classified as follow-up visits ranged from 19% to 23% and from 14% to 18%, respectively.
The NAMCS gathers data on a nationally representative sample of visits to office-based physicians who are primarily engaged in direct patient care. Each physician is randomly assigned to a 1-week reporting period. During this period, physicians or office staff record patient demographics, symptoms, procedures, diagnoses, and prescribed medications for a systematic random sample of visits.40 We included all visits to general or family medicine physicians, internists, pediatricians, and related specialties (Appendix). During the study period, physician participation rates in NAMCS ranged from 63% to 73%, and the unweighted item nonresponse rates were typically ≤5%.
The NHAMCS collects data on a nationally representative sample of visits to hospital-based emergency departments and outpatient clinics. Hospital staff record patient demographics, symptoms, procedures, diagnoses, prescribed medications, and hospital characteristics for a systematic random sample of patient visits during a randomly assigned 4-week reporting period.41 Because NHAMCS did not include a register of medical specialties, we included all visits to emergency departments and general medicine and pediatric outpatient clinics. The rates of participation in NHAMCS ranged from 85% to 97% for emergency departments and from 73% to 95% for outpatient clinics, whereas the unweighted item nonresponse rates were typically ≤5%.39
This study was approved by the institutional review boards of Vanderbilt University Medical Center and the Centers for Disease Control and Prevention.
As recommended by the National Center for Health Statistics, to obtain robust estimates, data from individual years were combined in pairs, and all of the reported estimates were based on >30 unweighted observations and relative SEs <30%.39 The study years were divided into a pre-PCV7 (1994–1995, 1996–1997, and 1998–1999), a transition (2000–2001), and a post-PCV7 (2002–2003) period. Denominators for rate calculations were the annual age-specific population estimates obtained from the US Census Bureau.42 For each 2-year period, the population was calculated by averaging the population on July 1 of the encompassing calendar years.
Because the protective effect of PCV7 against invasive pneumococcal diseases has been consistently observed among children aged <2 years who are the target of primary vaccination, we sought to determine whether pneumonia and otitis media rates decreased in children <2 years old relative to those aged 3 to 6 years.11,25,27,30–32 Children 3 to 6 years of age were less likely to receive PCV7 during the study period and, hence, were included as the reference group to account for temporal variability in visit rates for the study outcomes. Although variations in severity of winter viruses and secular trends in the diagnosis of otitis media and pneumonia could affect disease rates, we assumed that children in both study age groups would be similarly affected. Thus, rate ratios (RRs) of visits for the younger compared with the older age group should remain constant in the absence of vaccination. A decrease in the observed visit rate in children <2 years of age relative to the rate in those 3–6 years of age would be manifest as a decrease in the RRs after PCV7 licensure.31 Of note, any protective “herd” effect of the vaccine in older children43 would lead to an underestimation of the vaccine effect.
A Poisson regression model with 1 time term interaction was fit for each specific diagnosis group to obtain period RRs and variance estimates while accounting for the complex survey sampling design. The offset term in the models was represented by the population estimates calculated by specific time periods and age groups. Population estimates were distributed according to the aggregated weight factor for the surveys.44 The ratio of estimated RRs was used to assess changes over time periods. A negative binomial regression model for survey data using the same variables showed similar estimates suggesting no overdispersion. To quantify the potential effect of PCV7, expected visit rates for children aged <2 years in the transition (2000–2001) and post-PCV7 period (2002–2003) were calculated by multiplying the observed rates in children aged 3 to 6 years for those periods by the RR estimated from the pre-PCV7 period (1994–1999). Differences between calculated expected rates and the observed rates estimated the vaccine effect. The percentage change in rates was calculated, dividing the rate difference by the expected rate and multiplying by 100. The 95% confidence interval (CI) for the expected rate was computed using the delta method, which accounted for variances in both the RR and observed rates.45
During the pre-PCV7 period, average annual rates of otitis media visits declined from 1415 to 1072 visits per 1000 children aged <2 years and from 439 to 293 per 1000 children 3 to 6 years of age. The similar trends in these rates were reflected by a stable RR during the pre-PCV7 period (Fig 1). During the post-PCV7 period, we observed a significant decline in the RR as compared with the pre-PCV7 period (P = .014; Table 1), indicating a significant reduction in the rate of otitis media among children <2 years of age as compared with those aged 3 to 6 years. We observed an estimated increase of 119 (95% CI: −97 to 335) visits per 1000 children during the transition period and a significant decrease of 246 (95% CI: 25 to 466) otitis media visits per 1000 children <2 years old during the post-PCV7 period (Fig 2). Compared with the pre-PCV7 period, the rate of otitis media visits in children <2 years old increased 12% (95% CI: −10% to 35%) during the transition period and then decreased significantly, with a 20% (95% CI: 2% to 38%) decline in the post-PCV7 period (Fig 3).
Throughout the study period, children aged <2 years had an average of 74 to 92 combined pneumococcal and nonspecific pneumonia visits per 1000 children annually. Visit rates for children aged 3 to 6 years were consistently lower with 35 to 52 visits per 1000 children per year. The RRs for children <2 years old as compared with those 3 to 6 years of age trended upward during the pre-PCV7 period and downward after the PCV7 introduction (Fig 1). Nevertheless, the RR in the post-PCV7 period was not significantly different from the RR during the pre-PCV7 period (Table 1).
Using the pre-PCV7 average RR and observed visit rates in children 3 to 6 years to calculate expected rates in children aged < 2 years, 14 (95% CI: −50 to 77) and 35 (95% CI: −26 to 96) fewer pneumonia visits than expected per 1000 children <2 years old were observed during the transition and the post-PCV7 periods (Fig 2). However, the declines of 15% (95% CI: −53% to 83%) and 31% (95% CI: −23% to 85%) in pneumonia visit rates during the transition and post-PCV7 periods (Fig 3), respectively, were not statistically significant.
During the study period, rates of other ARIs ranged from 1074 to 1195 per 1000 for children <2 years of age. In children aged 3 to 6 years of age, average annual rates of other ARIs ranged from 521 to 595 per 1000 children. No significant variations in these rates were observed during the post-PCV7 period, and RRs were similar to those observed during the pre-PCV7 period (Fig 1). After PCV7 was licensed, rates of other ARIs in children <2 years old had an initial increase of 10% (95% CI: −10% to 30%) and then a decrease of 6% (95% CI: −10% to 22%; Fig 3). These fluctuations were not statistically significant (Table 1).
Randomized, controlled trials and observational studies have provided convincing support for the effectiveness of PCV7 in preventing invasive pneumococcal diseases in young children.11,25,27,28,30–32 However, the national impact of the PCV7 vaccination program on noninvasive pneumococcal-related diseases has not been assessed. Our study demonstrates a 20% decline in rates of otitis media visits in children aged <2 years after the introduction of PCV7 in the United States. Given the large societal and health care burden of otitis media, our findings suggest a positive national impact of the PCV7 vaccination program.
The significant decline in rates of otitis media in our study exceeded the effects reported in the randomized clinical trials.16,23,25,29,46 Although vaccine effectiveness in open populations has been reported to be greater than initially observed in controlled trials because of herd immunity,18,33,47,48 the magnitude of the decline that we observed was striking. However, using administrative databases and a similar ecological approach, Poehling et al31 reported similar findings with an overall 6% and 20% reduction in the rates of otitis media visits in children <2 years old during the post-PCV7 period in Tennessee Medicaid and New York private insurance populations, respectively.
In addition to herd immunity, several other factors may have influenced the observed trends in otitis media rates; these include more stringent diagnostic criteria for otitis media, lower rates of sick visits to health care providers, or lower diagnostic rates for otitis media when care is sought.49,50 However, the observed decline in rates of otitis media visits before the introduction of PCV7 affected both age groups similarly.31,50 Although influenza vaccine may prevent otitis media51 and vaccination for young children was first encouraged (when feasible) in 2002,52 the median national vaccine coverage remained <9% during the study period,53,54 and there was a mismatch between vaccine and circulating strains during the 2003–2004 influenza season.55 Interestingly, no significant changes in otitis media rates were observed during the transition period. This may be because the peak of otitis media incidence occurs between 6 and 18 months of age, and the impact of the vaccine may not be evident until children attain this age. In the Kaiser clinical trial, PCV7 was most effective against otitis media between ages 15 and 18 months shortly after a booster dose (12.2%; 95% CI: 8.2% to 16.0%), and the vaccine was more effective against the outcome “frequent otitis media” than a first episode of otitis media.26,29 Antibody titers required to protect against invasive pneumococcal diseases seem to be lower than those required to protect against otitis media,56 and the administration of a PCV7 booster dose (recommended by age 12–15 months1,24) could also contribute by reducing nasopharyngeal colonization by pneumococcal serotypes included in PCV7.48
The previously reported efficacy of PCV7 in preventing pneumonia in clinical trials varied depending on the definition of pneumonia. Clinical trial data consistently suggest a protective effect against radiologically confirmed pneumonia,11,26,28 however, the protection has been less impressive against simply clinically diagnosed cases.11,26,28 Clinically diagnosed pneumonia in these efficacy trials was comparable to our case definition, which was based on a coded diagnosis of pneumonia in the outpatient setting. Although our study did not detect decreases in pneumonia outpatient visits, the assessment of the full impact on pneumonia will require evaluation of pneumonia hospitalizations.
We performed an ecological study to evaluate the national impact of the PCV7 vaccination program on common noninvasive pneumococcal-related diseases. Ecological studies are considered the design of choice to evaluate the impact of policies when it is possible to rule out or control for secular trends. Because the policy impact is evaluated regardless of individual vaccination status, ecological studies can avoid selection bias.57 This assessment of outpatient medical care used publicly available national surveys that have been extensively reviewed and provide estimates with a reasonable degree of accuracy. By combining emergency department, office-based, and hospital-based clinic records, the entire spectrum of outpatient care was covered.
Several limitations of our study must be considered. For our ascertainment of noninvasive pneumococcal-related diseases, we relied on clinical diagnoses made by the physician. Radiologic information supporting the diagnosis of pneumonia was limited, and no distinction between clinical and radiologically confirmed pneumonias was made. Moreover, the survey data did not allow the distinction between initial and follow-up visits over the entire study period. The inclusion of follow-up visits may partially explain why the estimated pneumonia rates in our study were consistently higher than those reported previously. However, inclusion of all visits provides a reliable assessment of the overall outpatient health care use because of these pneumococcal-related conditions before and after the introduction of PCV7.4,5,9,58 Finally, these data excluded federally operated facilities and did not record information on consultations conducted over the telephone.
During the first years of licensure, there were documented vaccine shortages, delaying the distribution and administration of the vaccine. Thus, the vaccine effect observed may have been less than would have been seen with full implementation of the PCV7 vaccine recommendations. The Centers for Disease Control and Prevention recommended concentrating the vaccination efforts on infants and deferring vaccination for children >2 years of age, except for those at increased risk for pneumococcal diseases. However, the effect of this recommendation remains unknown.59,60 The National Immunization Survey estimated that by the end of the study period, 68.1% ± 1% of US children aged between 19 and 35 months had received ≥3 doses of PCV7, with considerable variability in uptake by state.61
Our results suggest a significant protective effect of the PCV7 vaccination program in the US outpatient setting with a substantial decline in otitis media visit rates. Given the remarkably high incidence of otitis media in children, its close association with the emergence of antibiotic resistance, and its huge economic burden, the persistence of the observed decline should herald a major change in the epidemiology of this disease and in our current clinical practice.
APPENDIX: ICD-9-CM DIAGNOSIS CODES AND NAMCS PHYSICIAN SPECIALTIES INCLUDED IN THE ANALYSIS.
ICD-9-CM Diagnosis Codes
Pneumonia and nonspecific pneumonias codes included 481, 482.9, 485, and 486.
Otitis media codes included 381, 381.0, 381.00, 381.01, 381.02, 381.03, 381.04, 381.05, 381.06, 381.1, 381.10, 381.19, 381.2, 381.20, 381.29, 381.3, 381.4, 381.5, 381.50, 381.51, 381.52, 381.6, 381.60, 381.61, 381.62, 381.63, 381.7, 381.8, 381.81, 381.89, 381.9, 382, 382.0, 382.00, 382.01, 382.02, 382.1, 382.2, 382.3, 382.4, and 382.9.
Other ARI codes included: 482, 482.0, 482.1, 482.2, 482.3, 482.30, 482.31, 482.32, 482.39, 482.4 482.40, 482.41, 482.49, 482.8, 482.81, 482.82, 482.83, 482.84, 482.89, 480, 480.0, 480.1, 480.2, 480.3, 480.8, 480.9, 487, 487.0, 487.1, 487.8, 466.1, 466.11, 466.19, 493.01, 493.02, 493.11, 493.12, 493.21, 493.22, 493.91, 493.92, 461, 461.0, 461.1, 461.2, 461.3, 461.8, 461.9, 466, 466.0, 491.21, 460, 462, 465.0, 463, 478.9, 493, 493.0, 493.00, 490, 464, 464.0, 464.00, 464.01, 464.1, 464.10, 464.11, 464.2, 464.20, 464.21, 464.3, 464.30, 464.31, 464.4, 464.50, 464.51, 465, 465.0, 465.8, 465.9, 483, 483.0, 483.1, 483.8, 484, 484.1, 484.3, 484.5, 484.6, 484.7, and 484.8.
NAMCS Physician Specialties
Family practice, geriatric medicine
Sports medicine (family practice)
Critical care pediatrics
Pediatric infectious diseases
Pediatric emergency medicine
Allergy and immunology
Adolescent medicine (internal medicine)
Critical care medicine
Sports medicine (emergency medicine)
Medical toxicology (emergency medicine)
General preventive medicine
Sports medicine (internal medicine),
Medical oncology pulmonary critical care medicine
Pediatric emergency medicine (emergency medicine)
Public health/general preventive medicine
This research study was funded in part by the Centers for Disease Control and Prevention New Vaccine Surveillance Network Cooperative Agreement U38/CCU417958 and an Association of Teachers of Preventive Medicine and Centers for Disease Control and Prevention Cooperative Agreement (TS-0825). Dr Poehling is supported by the Robert Wood Johnson Foundation Generalist Physician Faculty Scholar Program.
- Accepted April 19, 2006.
- Address correspondence to Marie R. Griffin, MD, MPH, Department of Preventive Medicine, Division of Pharmacoepidemiology, Vanderbilt University School of Medicine, A-1110 Medical Center North, 1161 21st Ave, Nashville, TN 37232-2637. E-mail:
Dr Nuorti and Ms Martin's current address is National Center for Immunization and Respiratory Diseases, Centers for Disease Control and Prevention, Atlanta, GA 30333
The authors have indicated they have no financial relationships relevant to this article to disclose.
- ↵Advisory Committee on Immunization Practices. Preventing pneumococcal disease among infants and young children. Recommendations of the Advisory Committee on Immunization Practices (ACIP). MMWR Recomm Rep.2000;49(RR-9) :1– 35
- Catterall JR. Streptococcus pneumoniae. Thorax.1999;54 :929– 937
- ↵Bluestone CD, Stephenson JS, Martin LM. Ten-year review of otitis media pathogens. Pediatr Infect Dis J.1992;11(8 suppl) :S7– S11
- ↵Murphy TF, Henderson FW, Clyde WA Jr, et al. Pneumonia: an eleven-year study in a pediatric practice. Am J Epidemiol.1981;113 :12– 21
- ↵Anderson RN, Smith BL. Deaths: leading causes for 2002. Natl Vital Stat Rep.2005;53(17) :1– 89
- ↵Kaplan B, Wandstrat TL, Cunningham JR. Overall cost in the treatment of otitis media. Pediatr Infect Dis J.1997;16(2 suppl) :S9– S11
- Gates GA. Cost-effectiveness considerations in otitis media treatment. Otolaryngol Head Neck Surg.1996;114 :525– 530
- ↵American Academy of Pediatrics, Subcommittee on Management of Acute Otitis Media. Diagnosis and management of acute otitis media. Pediatrics.2004;113 :1451– 1465
- ↵Auinger P, Lanphear BP, Kalkwarf HJ, et al. Trends in otitis media among children in the United States. Pediatrics.2003;112 :514– 520
- ↵American Academy of Pediatrics, Committee on Infectious Diseases. Policy statement: recommendations for the prevention of pneumococcal infections, including the use of pneumococcal conjugate vaccine (Prevnar), pneumococcal polysaccharide vaccine, and antibiotic prophylaxis. Pediatrics.2000;106 :362– 366
- ↵Poehling KA, Lafleur BJ, Szilagyi PG, et al. Population-based impact of pneumococcal conjugate vaccine in young children. Pediatrics.2004;114 :755– 761
- ↵Centers for Disease Control and Prevention. National Center for Health Statistics. Ambulatory health care data. Available at: www.cdc.gov/nchs/about/major/ahcd/ahcd1.htm. Accessed August 3, 2005
- ↵Centers for Disease Control and Prevention. National Center for Health Statistics. Ambulatory health care data. Available at: www.cdc.gov/nchs/about/major/ahcd/namcsdes.htm. Accessed August 2, 2005
- ↵Centers for Disease Control and Prevention. National Center for Health Statistics. Ambulatory health care data. Available at: www.cdc.gov/nchs/about/major/ahcd/nhamcsds.htm. Accessed August 2, 2005
- ↵United States Census Bureau. Population estimates datasets. Available at: www.census.gov/popest/datasets.html. Accessed August 23, 2005
- ↵Casella G, Berger RL. Statistical Inference. 2nd ed. Duxbury, CA: Duxbury Press; 2002
- ↵Kilpi T, Ahman H, Jokinen J, et al. Protective efficacy of a second pneumococcal conjugate vaccine against pneumococcal acute otitis media in infants and children: randomized, controlled trial of a 7-valent pneumococcal polysaccharide-meningococcal outer membrane protein complex conjugate vaccine in 1666 children. Clin Infect Dis.2003;37 :1155– 1164
- ↵Halloran ME, Haber M, Longini IM Jr, et al. Direct and indirect effects in vaccine efficacy and effectiveness. Am J Epidemiol.1991;133 :323– 331
- ↵Ghaffar F, Barton T, Lozano J, et al. Effect of the 7-valent pneumococcal conjugate vaccine on nasopharyngeal colonization by Streptococcus pneumoniae in the first 2 years of life. Clin Infect Dis.2004;39 :930– 938
- ↵Finkelstein JA, Stille C, Nordin J, et al. Reduction in antibiotic use among US children, 1996–2000. Pediatrics.2003;112 :620– 627
- ↵Harper SA, Fukuda K, Uyeki TM et al. Prevention and control of influenza. Recommendations of the Advisory Committee on Immunization Practices (ACIP) [published correction appears in MMWR Morb Mortal Wkly Rep. 2005;54:750]. MMWR Recomm Rep.2005;54(RR-8) :1– 40
- ↵Bridges CB, Fukuda K, Uyeki TM, et al. Prevention and control of influenza. Recommendations of the Advisory Committee on Immunization Practices (ACIP). MMWR Recomm Rep.2002;51(RR-3) :1– 31
- ↵Jokinen C, Heiskanen L, Juvonen H, et al. Incidence of community-acquired pneumonia in the population of four municipalities in eastern Finland. Am J Epidemiol.1993;137 :977– 988
- ↵Centers for Disease Control and Prevention. Decreased availability of Pneumococcal Conjugate Vaccine. MMWR Morb Mortal Wkly Rep.2001;50 :783– 784
- ↵Centers for Disease Control and Prevention. National, state, and urban area vaccination coverage among children aged 19–35 months: United States, 2004. MMWR Morb Mortal Wkly Rep.2005;54 :718– 724
- Copyright © 2006 by the American Academy of Pediatrics