Pertussis infection, also called whooping cough, is a highly contagious bacterial disease caused by Bordetella pertussis. Symptoms include uncontrollable violent coughing, particularly in infants. Pertussis vaccines are well established and recommended by the World Health Organization,1 yet pertussis was still one of the leading causes of vaccine-preventable deaths worldwide in children aged <5 years in 2008.2
In the US, 48,277 cases of pertussis were reported in 2012.3 The highest incidence of pertussis was in infants aged <1 year, who are at the greatest risk for severe disease and/or death.3 Of the 20 pertussis deaths reported in the US in 2012, 16 (80%) were in infants aged <1 year, and 15 (75%) were in infants aged <3 months.3
Pertussis vaccination programs have substantially reduced pertussis incidence in industrialized countries.1 However, a vaccination program effectiveness depends on compliance with the recommended schedule. In the US, the Advisory Committee on Immunization Practices (ACIP) recommended diphtheria– tetanus–pertussis (DTaP) schedule includes a primary series of 3 doses at age 2, 4 and 6 months (with a time window of 1 month for each dose), followed by a fourth dose at age 18 months and a later booster dose.4 Vaccine coverage levels are high in the US, with 84.6% of children aged 19–35 months receiving ≥4 doses of DTaP vaccine in 2011.5 However, high coverage can mask substantial delays before full vaccination. In 2003, children were undervaccinated for a mean of 172 days for all vaccines during their first 24 months of life.6 Approximately 34% were undervaccinated for <1 month, 29% for 1–6 months and 37% for >6 months.6 In Australia, although DTaP coverage rates increased from 88% to 92% between 1998 and 2001, the vaccination delay remained unchanged and the authors concluded that measurement of timeliness should be incorporated into routine monitoring of immunization program delivery success in countries with high coverage levels.7
Studies in the US,6,8 Australia,7 Sweden9 and the Netherlands10 have reported delays in vaccine administration that would leave children temporarily with lower levels of protection than anticipated. Vaccination timeliness is important for several reasons. In the US, children who fell behind in their vaccinations were less likely to be fully vaccinated later on.6 Delays may also interrupt the vaccine schedule, increasing the number of office visits the parent must make to the provider, which can adversely affect compliance.11 Prompt vaccination decreases the risk not only to the child, which may be especially important for infections such as pertussis that are potentially severe in young infants,7 but may also reduce the risk of disease in the wider community.11
The primary objective of this study was to estimate DTaP vaccination timeliness for the first 3 doses and investigate the potential impact of improved timeliness on health outcomes and costs in the US, by comparing the current situation with a hypothetical situation of delivery of the 3 primary DTaP doses at exactly 60, 120 and 180 days of age. In addition, a secondary objective was to identify which factors were associated with vaccination delay.
MATERIALS AND METHODS
Current DTaP Vaccine Timeliness in the US
The DTaP infant vaccination schedule recommended in the US by ACIP is for administration of the first 3 doses at age 2, 4 and 6 months.4 Thus, the time windows begin at the age of 60, 120 and 180 days, respectively.
To estimate current DTaP vaccine timeliness in the US, we analyzed data from the 2010 National Immunization Survey (NIS),12 the most recent information available when the analysis was conducted. This survey was conducted by the Centers for Disease Control and Prevention to estimate vaccination timeliness for US children aged ≤35 months. We used data for children aged ≤12 months. Approximately 17,000 subjects with available data were included. The NIS uses random digit dialing telephone survey methodology to identify households containing children in the target age range and interviews the adult most knowledgeable about the child’s vaccinations. With consent of the child’s parent or guardian, the NIS contacts (by mail) the child’s health care provider(s) to request vaccination information from the child’s medical records. The NIS uses the synthesized provider-reported vaccination histories to estimate vaccination coverage because they are considered more accurate than household-reported data. Each child is weighted in the database using a sampling weight, used in the analysis to ensure that the results represent the US population. The analysis was based on children with available information at each time point.
We used the NIS dataset to calculate the mean age at DTaP vaccination for each dose. Individuals who did not receive the dose were not included in the calculation of mean age at that vaccination.
A static Markov cohort model has previously been developed to estimate the health outcomes and costs associated with various pertussis vaccination strategies.13 In this analysis, we considered only the infant vaccination strategy. The model was initially developed for the Netherlands and adapted for the US by entering country-specific incidence and cost data in Microsoft Excel. The model was run using the 2009 US birth cohort of 4,131,019 infants,14 followed for the first year of life. As the time horizon was a single year, costs were not discounted. For a death caused by pertussis in the 1-year time horizon, the lifetime loss in quality-adjusted life years (QALYs) was estimated, discounted at 3% per year. The cost analysis was conducted from the perspective of the US healthcare system including only direct pertussis-related medical costs. Indirect costs because of lost productivity were not included. Costs associated with ensuring timely vaccination (eg, reminder calls, e-mails, texts) were not taken into account. All costs were in 2012 US dollar ($). The modelled outcomes included health outcomes (number of pertussis cases, hospitalizations, deaths and QALYs lost) and total direct medical costs of pertussis illness.
Data on the age-specific incidence (for month of life) of pertussis and the probabilities of associated hospitalization and death for the US during 2000–2007 were taken from a recently published US DTaP vaccination cost-effectiveness study.15 These data were based on data from the US National Notifiable Disease Surveillance System, with an assumption of 15% underreporting.15 Underreporting in this context refers to cases that were not identified as pertussis on the surveillance system. These unreported cases may have been hospitalized or attended a medical practitioner but were categorized as something other than pertussis and therefore not identified on the surveillance system. The estimate of 15% unreported cases used in the present analysis may be conservative as underreporting rates of up to 50% in infants in the US have been reported elsewhere.16
Vaccination coverage was set at 96.5% for 1 dose, 94.8% for 2 doses and 89.8% for 3 doses, calculated from the NIS survey data for the first 12 months.12 Vaccine effectiveness was assumed to be 40% after the first dose, 70% after the second dose and 89% after the third dose, extrapolated from Dutch incidence data and follow-up studies in the US17 and Sweden18 and used in a previous analysis.13
Direct medical costs per case in 2012 were as follows: outpatient treatment for respiratory complications $127, hospitalization for respiratory complications $11,205, neurological complications $8101 and death $18,212.19 Unreported cases were assumed to incur no medical costs.
The mean duration of disease for infants was 80 days.20 Disutility for hospitalized infants was 0.42 and 0.15 for mild to moderate/unreported case (not hospitalized), based on published data.20 These disutility data were also used in a previous analysis.13 It is expected that all cases of pertussis in infants will be medically attended but not all may be reported as pertussis in surveillance systems. Thus, applying the disutility for nonhospitalized pertussis to all unreported cases is likely to be a conservative assumption, as some unreported cases may actually be hospitalized.16 The annual QALY loss per case, taking into account the disutility and the duration of illness, ranged from 0.033 for a nonhospitalized case to 0.092 for a hospitalized case.
Data on the observed incidence of pertussis reflect the effects of pertussis vaccination at current timeliness. To estimate the impact of a hypothetical situation with vaccination at exactly 60, 120 and 180 days, we used a 2-step approach. First, we estimated the number of cases expected with no vaccination (NV) by removing the overall effect of current (C) vaccine coverage (Cov) and efficacy (Eff) from the model as follows:
The formula was adjusted for the coverage, efficacy and average date for each of the 3 primary vaccine doses. Second, we ran the model with the same vaccine coverage and efficacy but with the 3 doses administered at exactly 60, 120 and 180 days. This estimated the incidence expected in a hypothetical situation of vaccination at these times. Comparing these results with the results estimated for current vaccination timeliness provided an estimate of the expected potential impact of timely vaccination.
A linear regression analysis was conducted, using the regression procedure PROC Reg in SAS (Version 9.2), to explore the factors associated with vaccination delay. The age (in days) of the child at the third vaccination dose was used as the dependent (outcome) variable. The NIS dataset includes information on family and child characteristics, financial situation and insurance cover.12 These variables were introduced in univariate and multivariate regression analysis. A significance level of 0.001 was used to reduce the issue of multiple comparisons. A backward elimination process was used to generate the final multivariate model.
Current Timeliness of DTaP Vaccination in the US
Figure 1 shows the cumulative probability of vaccination by age in days for each of the 3 DTaP vaccine doses, from NIS data.
The schedule recommended by ACIP is for 3 primary doses to be administered at age 2, 4 and 6 months, with a time window of 1 month for each dose beginning at age 60, 120 and 180 days, respectively. The observed mean age at vaccination was 76 days for the first dose, 147 days for the second dose and 224 days for the third dose. These correspond to mean delays, relative to administration at exactly 60, 120 and 180 days, of 16, 27 and 44 days, respectively. Individuals who did not receive vaccination were not included in the calculation of mean age at vaccination.
Estimated Health Outcomes and Costs
Table 1 shows the modelled estimates for the number of pertussis cases, hospitalizations and deaths and the number of QALYs in the cohort during the first year of life.
Data are presented for current timeliness and the hypothetical situation of vaccination at exactly 60, 120 and 180 days. Figure 2 shows the distribution of the estimated pertussis cases and hospitalizations by age in months.
Timely vaccination would be expected to avoid 278 pertussis cases, 103 pertussis hospitalizations and 1 death and to reduce the number of QALYs lost to pertussis illness by 38.
The cost of the hospitalizations, neurological complications, deaths and outpatient visits avoided by the hypothetical situation of vaccination at exactly 60, 120 and 180 days was estimated at $1.03 million over the 1-year modelled time horizon.
The risk factors that were statistically significantly associated with delay in receiving the third vaccine dose in the univariate analysis are shown in Table 2. Other variables, analyzed but not significant in the univariate analysis were duration of breast feeding in days, Hispanic origin of child, language in which interview was conducted, number of working cell phones household members have available for personal use, number of residential telephone numbers in household (excluding cell phones), number of providers responding with vaccination data for child and gender of the child.
In the multivariate analysis (Table 3), risk factors significantly associated with vaccine delay included having a young, unmarried mother with low education and more than 1 child; a non-Hispanic black or mixed-race child; being eligible for the Vaccines For Children (VFC) program; having moved state and the geographical state of residence.
This analysis of national surveillance data showed that administration of DTaP vaccination to children aged <1 year in the US is delayed by an average of 16 days for the first dose, 27 days for the second dose and 44 days for the third dose, relative to vaccination at exactly age 60, 120 and 180 days. Vaccination delays lengthen the time during which young infants have reduced protection from pertussis. This may have serious consequences, as pertussis infection in young infants can be severe or fatal.3 Furthermore, delays in vaccination and the consequent presence of susceptible infants may enhance disease transmission and increase the risk of disease outbreaks. The results presented here estimated that a hypothetical situation of administration of the first 3 vaccine doses at exactly age 60, 120 and 180 days could potentially prevent 9.1% of pertussis cases, 7.2% of hospitalizations, 3.1% of deaths per year in infants aged <1 year in the US, compared with the current situation. This could also save an estimated $1.03 million per year in direct medical costs and gain an estimated 38 QALYs per year.
This study builds on a similar hypothetical exercise conducted in Sweden,9 by extending the analysis to cover estimates of the impact on costs and QALY gains. Greater benefits of timely vaccination were observed in the study in Sweden, which estimated a reduction of 28% in pertussis cases and 38% in hospitalizations,9 suggesting that our estimates may be conservative. The difference may reflect the finding in the Swedish study of a reduction in complications after the first vaccine dose.9 This study did not attempt to include any additional benefit of reduced disease severity after the first dose because of limited data availability to populate the model. This conservative assumption may have resulted in an underestimate of the effect of timely vaccination relative to the Swedish study and is a limitation of the present analysis, as noted below.
We used data on disutility and duration of illness from a US study,20 as our analysis related to the US. A more recent study using the EuroQoL 5-dimension questionnaire to evaluate quality of life in 535 patients aged 5 years or more with laboratory-confirmed pertussis in England and Wales estimated that the overall QALY loss for a laboratory-confirmed pertussis case was 0.097 QALY.21 This is similar to the QALY loss of 0.092 used for a hospitalized infant pertussis case in this study. However, the QALY loss in an infant hospitalized with pertussis would be expected to be considerably higher than the value reported in school-age children and adults in the survey, as pertussis is known to be more severe in infants than in older children and adults.21 This suggests that the QALY loss used in the present analysis may be conservative.
Timely vaccination is likely to be particularly important for diseases that circulate continually, such as pertussis, and diseases that can potentially cause large outbreaks. Poor vaccination coverage and delays in diseases such as rotavirus, measles and pertussis potentially increase the risk of epidemics, and timely vaccination is, therefore, particularly beneficial. In addition, timely vaccination is likely to be of particular benefit in pertussis because of the severity of pertussis infection in young infants. This is reflected in the low utility value of 0.58 associated with pertussis infection in hospitalized infants,20,22 and in the disproportionate concentration of pertussis deaths in young infants.3 The results presented here suggest that the dosing windows for timely vaccination should be stricter.
Improving the timeliness of vaccination for 1 disease could possibly have beneficial effects on others. When rotavirus vaccination was introduced in Australia in 2007, for the first time the Australian childhood vaccination schedule included a vaccine that required administration within narrow dosing windows. After rotavirus vaccination introduction, timeliness of the third dose of DTaP increased by 5–12 percentage points compared with the previous situation, suggesting that the inclusion of a vaccine with a strict dosing schedule may have encouraged parents to present promptly, with consequent improvement in other concurrently administered vaccines.23
Childhood vaccination schedules have become increasingly crowded as the number of diseases with recommended vaccines has grown.24 Approximately 15 vaccinations are recommended by the age of 19 months in the US.8 The complexity of the schedule and the number of separate doses required at each appointment can be intimidating and may contribute to vaccination delays. Combination vaccines reduce the number of doses required at each visit, compared with monovalent vaccines, and could help improve compliance. In Germany, vaccination timeliness improved as monovalent vaccines was replaced by quadrivalent, pentavalent and hexavalent vaccines.25 In a retrospective matched-cohort study in managed care in the US, children receiving the pentavalent DTaP/hepatitis B/inactivated polio vaccine (HepB/IPV) combination vaccine were significantly more likely to be fully vaccinated by age 2 years and to be vaccinated within the recommended age ranges, compared with children who did not receive the pentavalent vaccine.26
In multivariate analysis, factors associated with vaccination delay included having a young, unmarried mother without a college degree; a non-Hispanic black or mixed-race child; more than 1 child in the household and being eligible for the Vaccines For Children program, indicating poor financial status. These findings are consistent with previous studies that have reported insurance status, race, maternal age, education and marital status, large family size and Medicaid enrolment as predictors of vaccine delay.6,27–29
Our study has a number of limitations. First, recent pertussis incidence data, broken down by monthly age groups in children younger than 12 months, in the US are not available. As incidence rates are increasing in children younger than 1 year, it is likely that we have underestimated the impact of timely vaccination. The model did not include indirect costs, such as lost productivity if a parent has to take time away from work or usual activities to care for a child with pertussis, or the indirect cost of an infant death. As the indirect cost of an infant death has been estimated at $5.9 million,19 this limitation could considerably underestimate the benefits of improved vaccination timeliness.
Also, our study is based on a static model13 and thus cannot take account of herd protection effects. Little is known about transmission patterns in very young infants because studies often do not discriminate between infants, toddlers and children approaching school age. For example, a large study of social contact and mixing patterns in Europe had only 1 age group covering infants and children aged 0–4 years.30 Clearly, the contact pattern of an infant aged <6 months would be expected to differ from that of a toddler or a child attending kindergarten. Studies refining the contact matrix for very young infants would help to improve understanding of infectious disease transmission in this age group and would be a useful area for future research.
A further limitation of the model is that it could underestimate any benefits of reductions in disease severity obtained after partial vaccination. Recent publications suggest that children with partial vaccination (eg, 1–2 doses) have reduced disease severity if they develop pertussis disease.9,31–33 Also, unreported pertussis cases may be hospitalized but not recognized as pertussis cases, and thus, the assumption that unreported cases incur no medical costs may be conservative. However, costs associated with targeted strategies to improve timeliness of vaccination (eg, reminder calls, e-mails, texts) were not taken into account.
Our results indicate that improving the timeliness of pertussis vaccination could potentially save $1.03 million over 1 year in direct medical costs in the US. Reinvestment of such future savings into the development of interventions to improve timeliness of vaccination in high-risk groups could offer valuable public health benefits. For example, a community-wide reminder, recall and outreach program successfully reduced disparities in immunization rates between inner-city and suburban populations and between ethnic groups in the area of Rochester, New York.34 In another US study, targeted text message reminders were associated with increased influenza vaccination of children and adolescents in a low-income urban population.35 Such targeted interventions could help to improve pertussis vaccination timeliness, helping to realize the potential benefits suggested by this study.
Administration of the first 3 pertussis vaccine doses to infants at exactly 60, 120 and 180 days of age could avoid 278 pertussis cases, 103 hospitalizations and 1 death per year in the US, with estimated savings of $1.03 million per year in direct medical costs. Future savings in medical costs could be redirected to develop interventions to help improve vaccine timeliness in groups at high risk for delayed vaccination.
The authors would like to thank Carole Nadin (Fleetwith Ltd., Ipswich, United Kingdom on behalf of GSK Vaccines, Wavre, Belgium) who provided medical writing services, as well as Jennifer Dorts, Bruno Baudoux and Nathalie Arts (Business and Decision Life Sciences, Brussels, Belgium on behalf of GSK Vaccines, Wavre, Belgium), who provided publication coordination.
Author contributions: All authors conceived and designed the study; A.T. developed the model with the input of all authors. The data were acquired by A.T. and analyzed by all the authors; A.T., C.M.and J.-E.P. conducted the literature review. All authors reviewed and commented on manuscript drafts. All authors had full access to the data and gave final approval before submission.
1. World Health Organization. . Pertussis
vaccines: WHO position paper. Wkly Epidemiol Rec. 2010;85:385–400
2. World Health Organization. . Estimates of disease burden, 2008 2014Accessed November 14, 2014 Available at: http://www.who.int/immunization/monitoring_surveillance/burden/estimates/en/index.html
3. Centers for Disease Control and Prevention. . Pertussis
: surveillance and reporting. 2013Accessed January 9, 20142012 Final pertussis
surveillance report Available at: http://www.cdc.gov/pertussis/downloads/pertuss-surv-report-2012.pdf
4. Akinsanya-Beysolow I, Jenkins R, Meissner HC. Advisory Committee on Immunization Practices (ACIP) recommended immunization schedule for persons aged 0 through 18 years—United States, 2013. MMWR Surveill Summ. 2013;62(Suppl 1):2–8
5. Centers for Disease Control and Prevention. . National, state, and local area vaccination
coverage among children aged 19–35 months—United States, 2011. MMWR Morb Mortal Wkly Rep. 2012;61:689–696
6. Luman ET, Barker LE, Shaw KM, et al. Timeliness
of childhood vaccinations in the United States: days undervaccinated and number of vaccines delayed. JAMA. 2005;293:1204–1211
7. Hull BP, McIntyre PB. Timeliness
of childhood immunisation in Australia. Vaccine. 2006;24:4403–4408
8. Luman ET, Barker LE, McCauley MM, et al. Timeliness
of childhood immunizations: a state-specific analysis. Am J Public Health. 2005;95:1367–1374
9. Nilsson L, Lepp T, von Segebaden K, et al. Pertussis vaccination
in infancy lowers the incidence of pertussis
disease and the rate of hospitalisation after one and two doses: analyses of 10 years of pertussis
surveillance. Vaccine. 2012;30:3239–3247
10. van der Ploeg CP, Kateman H, Vermeer-de Bondt PE, et al. [Increased risk of hepatitis B due to incomplete or untimely immunisation in one-quarter of infants of hepatitis-B-virus carriers]. Ned Tijdschr Geneeskd. 2004;148:1820–1824
11. Koslap-Petraco MB, Parsons T. Communicating the benefits of combination vaccines to parents and health care providers. J Pediatr Health Care. 2003;17:53–57
12. Centers for Disease Control and Prevention. . National Immunization Survey 2010 2010Accessed June 24, 2015 Available at: http://www.cdc.gov/nchs/nis/data_files.htm
13. Westra TA, de Vries R, Tamminga JJ, et al. Cost
-effectiveness analysis of various pertussis vaccination
strategies primarily aimed at protecting infants in the Netherlands. Clin Ther. 2010;32:1479–1495
14. Hamilton BE, Martin JA, Ventura SJ. Births: preliminary data for 2009. Natl Vital Stat Rep. 2010;59:1–19
15. Terranella A, Asay GR, Messonnier ML, et al. Pregnancy dose Tdap and postpartum cocooning to prevent infant pertussis
: a decision analysis. Pediatrics. 2013;131:e1748–e1756
16. Cortese MM, Baughman AL, Zhang R, et al. Pertussis
hospitalizations among infants in the United States, 1993 to 2004. Pediatrics. 2008;121:484–492
17. Bisgard KM, Rhodes P, Connelly BL, et al.Centers for Disease Control and Prevention, Atlanta, Georgia 30333, USA. firstname.lastname@example.org
vaccine effectiveness among children 6 to 59 months of age in the United States, 1998-2001. Pediatrics. 2005;116:e285–e294
18. Gustafsson L, Hessel L, Storsaeter J, et al. Long-term follow-up of Swedish children vaccinated with acellular pertussis
vaccines at 3, 5, and 12 months of age indicates the need for a booster dose at 5 to 7 years of age. Pediatrics. 2006;118:978–984
19. Ding Y, Yeh SH, Mink CA, et al. Cost
-benefit analysis of hospital based postpartum vaccination
with combined tetanus toxoid, reduced diphtheria toxoid, and acellular pertussis
vaccine (Tdap). Vaccine. 2013;31:2558–2564
20. Lee GM, Murphy TV, Lett S, et al. Cost
effectiveness of pertussis vaccination
in adults. Am J Prev Med. 2007;32:186–193
21. van Hoek AJ, Campbell H, Andrews N, et al. The burden of disease and health care use among pertussis
cases in school aged children and adults in England and Wales; a patient survey. PLoS One. 2014;9:e111807
22. Lee GM, Salomon JA, LeBaron CW, et al. Health-state valuations for pertussis
: methods for valuing short-term health states. Health Qual Life Outcomes. 2005;3:17
23. Wendy B. Vaccination
with 3-dose paediatric rotavirus vaccine (RotaTeq®): impact on the timeliness
of uptake of the primary course of DTPa vaccine. Vaccine. 2012;30:5293–5297
24. Zepp F, Schmitt HJ, Cleerbout J, et al. Review of 8 years of experience with Infanrix hexa (DTPa-HBV-IPV/Hib hexavalent vaccine). Expert Rev Vaccines. 2009;8:663–678
25. Kalies H, Grote V, Verstraeten T, et al. The use of combination vaccines has improved timeliness
in children. Pediatr Infect Dis J. 2006;25:507–512
26. Happe LE, Lunacsek OE, Marshall GS, et al. Combination vaccine use and vaccination
quality in a managed care population. Am J Manag Care. 2007;13:506–512
27. Dombkowski KJ, Lantz PM, Freed GL. Risk factors for delay in age-appropriate vaccination
. Public Health Rep. 2004;119:144–155
28. Fiks AG, Alessandrini EA, Luberti AA, et al. Identifying factors predicting immunization delay for children followed in an urban primary care network using an electronic health record. Pediatrics. 2006;118:e1680–e1686
29. Strine TW, Luman ET, Okoro CA, et al. Predictors of age-appropriate receipt of DTaP dose 4. Am J Prev Med. 2003;25:45–49
30. Mossong J, Hens N, Jit M, et al. Social contacts and mixing patterns relevant to the spread of infectious diseases. PLoS Med. 2008;5:e74
31. Quinn HE, Snelling TL, Macartney KK, et al. Duration of protection after first dose of acellular pertussis
vaccine in infants. Pediatrics. 2014;133:e513–e519
32. Sheridan SL, McCall BJ, Davis CA, et al. Acellular pertussis
vaccine effectiveness for children during the 2009-2010 pertussis
epidemic in Queensland. Med J Aust. 2014;200:334–338
33. Barlow RS, Reynolds LE, Cieslak PR, et al. Vaccinated children and adolescents with pertussis
infections experience reduced illness severity and duration, Oregon, 2010-2012. Clin Infect Dis. 2014;58:1523–1529
34. Szilagyi PG, Schaffer S, Shone L, et al. Reducing geographic, racial, and ethnic disparities in childhood immunization rates by using reminder/recall interventions in urban primary care practices. Pediatrics. 2002;110:e58
35. Stockwell MS, Kharbanda EO, Martinez RA, et al. Effect of a text messaging intervention on influenza vaccination
in an urban, low-income pediatric
and adolescent population: a randomized controlled trial. JAMA. 2012;307:1702–1708