Background
Streptococcus pneumoniae (
S. pneumoniae) is the single most significant bacterial cause of invasive (meningitis and bacteremia) and noninvasive (pneumonia and otitis media) diseases in children <5 years of age worldwide [
1]. In China, pneumococcal diseases pose a major burden in young children and the elderly, leading to nearly 30,000 deaths annually [
2‐
4]. Pneumococcal diseases caused by
S. pneumoniae are primarily treated with penicillin; however, the increasing prevalence of drug-resistant pneumococci is a concern worldwide and demands more sophisticated disease management. The treatment of infections caused by antibiotic-resistant
S. pneumoniae would require more resources, including more expensive antibiotic agents and longer hospital stays, which would substantially increase healthcare expenditures [
5‐
7].
New pneumococcal conjugate vaccines have been developed to protect infants and young children from pneumococcal diseases [
8,
9]. The heptavalent pneumococcal conjugate vaccine (PCV-7) Prevnar
® (Pfizer Vaccines) was the first such vaccine licensed by the U.S. Food and Drug Administration, and it has been available since 2000. PCV-7 is composed of seven saccharides from the capsular antigen of
S. pneumoniae, each conjugated to a CRM
197 protein, which is a nontoxic mutant of diphtheria toxin carrier. PCV-7 contains the serotypes 4, 6B, 9V, 14, 18C, 19F, and 23F, which cause the majority of cases of invasive pneumococcal disease (IPD) worldwide [
10‐
12]. The clinical effectiveness and safety of PCV-7 in infants and children have been evaluated in several large-scale clinical trials. The results demonstrate that vaccination significantly decreased the incidence rates of IPD, pneumonia, and otitis media [
13‐
15]. Furthermore, surveillance data indicated that PCV-7 could reduce the incidence rate of pneumococcal disease in both vaccinated infants and in unvaccinated infants by “herd immunity”. The direct effect of the vaccine could also extend the benefit to adults by decreasing the nasopharyngeal presence of
S. pneumoniae in vaccinated children [
16‐
18]. At present, new vaccines of higher valence, namely, a 10-valent vaccine (PCV-10) called Synflorix® (GlaxoSmithKline) and a 13-valent vaccine (PCV-13) (Pfizer Vaccines), have been approved to replace PCV-7 in the future.
Following recommendations from the Strategic Advisory Group of Experts on Immunization, the World Health Organization (WHO) suggests that the pneumococcal conjugate vaccine should be included in national immunization programs to reduce the heavy pneumococcal disease burden [
19]. Over the last decade, many developed countries have added a pneumococcal vaccine to compulsory, routine immunization schedules [
20]. In poor countries, multilateral organizations, including the Global Alliance for Vaccines and Immunization (GAVI), have taken an active role in improving access to vaccines [
21]. At present, PCV-7 is the only pneumococcal conjugate vaccine licensed for use in infants and young children in China; however, the introduction of PCV-7 into the Chinese national immunization program still poses a challenge due to the high cost of the imported vaccine. The economic outcome of public health interventions is an important factor in policy decisions. Although economic outcomes in developed countries indicate that a universal PCV-7 vaccination program could reduce both medical and nonmedical costs, there is little economic evidence to support the universal use of this vaccine in China [
22]. The aim of the present study was to evaluate the projected health benefits, costs, and cost-effectiveness of universal infant vaccination with PCV-7 in China.
Discussion
Although the incidence rate in China is lower than in other developing regions, such as Africa and Southeast Asia, China accounts for 12% of pneumococcal cases worldwide because it has the largest population [
31]. The introduction of PCV vaccines has increased interest in preventing pneumococcal disease. Our results show that the routine compulsory vaccination with PCV-7 of Chinese infants has the potential to significantly decrease the disease burden and mortality associated with
S. pneumoniae by 33.8 and 57.4%, respectively. The surveillance data from the regions where PCV-7 has already been incorporated into the national immunization program demonstrate the effectiveness of this vaccine [
16,
55]. In the United States, a 75% decrease in the all-cause IPD incidence rate was observed within 3 years of vaccine introduction, and nearly all IPD caused by vaccine serotypes was prevented in children ≤5 years of age [
55]. The estimated cost associated with pneumococcal disease was reduced by 34.1% in the birth cohort receiving PCV-7 vaccination. However, our economic analysis indicated that the PCV-7 vaccination program in China is not cost-effective because the incremental cost per QALY gained is far greater than three times the per-capita GDP of China, which is the standard cutoff cost recommended by the WHO Choosing Interventions that are Cost Effective (WHO-CHOICE) project. This result persisted even when a 75% discount rate was applied to the per-dose price of PCV-7 [
56,
57]. To the best of our knowledge, this is the first cost-effectiveness evaluation of the economic outcome of PCV-7 vaccination implementation in China. The results may be helpful for Chinese health-policy makers in deciding whether to add PCV-7 to the compulsory routine vaccination program in China.
Our results indicated that the cost of the vaccine itself drives the cost of PCV-7 vaccination, accounting for 72.2 or 73.6% of the total cost with or without herd immunity, respectively. The one-way sensitivity analysis showed that varying the cost of PCV-7 from 75 to 125% of the base-case value yielded the most influential parameters. The literature has shown that the cost of PCV-7 vaccine acquisition in the rest of the world is substantially different from the cost in China. For example, the estimated current cost of PCV-7 per dose in the European national immunization program was nearly $80, which is 41% lower than the cost in China; [
58] however, only a 75% reduction in vaccine cost would yield a nearly 75% probability of cost-effectiveness when herd immunity was considered (Figure
4). To satisfy the standard of cost-effectiveness, Figure
3 suggests that the cost of PCV-7 should be reduced to at least 35.1 or 21.5% of the base-case value with or without herd immunity, respectively. This suggestion is in accordance with the study by Mari Nakamura
et al., which indicated that vaccination with PCV-7 would be cost-effective at a per-dose cost of $10 for lower-middle-income (2008 GNI per capita: $976–3,855) and $20 for upper-middle-income (2008 GNI per capita: $3,856–11,905) countries [
59]. The costs of introducing a compulsory routine pneumococcal vaccination program in China should be weighed against the losses based on more information regarding budget impact, affordability and sustainability.
Our findings project a 38% reduction in IPD following PCV-7 vaccination in China, which is greater than the reduction in pneumococcal AOM (32%); however, our results indicate that the major contributor to the cost savings from the introduction of PCV-7 vaccination emerges from a reduction in costs related to pneumococcal AOM. Otitis media has relatively less serious outcomes and a higher incidence rate compared with IPD according to this analysis, in agreement with the findings of previous studies. Pneumococcal non-typeable
Haemophilus influenza protein D conjugate vaccine (PHiD-CV) is a newly licensed 10-valent pneumococcal conjugate vaccine. It may protect against diseases caused by non-typeable
H. influenzae (NTHi) [
60]. Vaccination with PHiD-CV is projected to lead to a 33.6% decrease in the overall incidence of otitis media and a 35.6% reduction in the incidence of otitis media caused by
H. influenzae[
60]. It is expected that a more favorable ICER would be achieved if the cost of PHiD-CV were comparable to that of PCV-7. The base-case analysis also showed that the cost associated with pneumococcal pneumonia was the second major contributor to the total economic burden of pneumococcal disease. According to a recent study by Kai-Hu Yao
et al., serotype 19A, which is not covered by PCV-7, is the second most common serotype in Chinese children [
36]. The serotype coverage rate of PCV-13, which does include 19A, for pneumococcal pneumonia is 16% higher than that of PCV-7. New pneumococcal vaccines with higher serotype coverage rates should be given serious consideration when deciding whether to support pneumococcal vaccination in China. The current analysis should be updated when PCV-10, PCV-13 and PHiD-CV are supplied in China.
Previous studies have shown that the addition of indirect effects offer considerable support for the universal vaccination of young children [
43]. Although efforts to introduce PCVs in developing countries are increasing, few published reports have used local data to account for the role of herd immunity from universal childhood PCV-7 vaccination in reducing adult cases of IPD. The evidence from early epidemiological studies showed that herd immunity decreases the overall burden of pneumococcal disease [
41,
55]. In comparison to a model that does not account for herd immunity from universal infant PCV-7 vaccination in China in terms of reducing adult cases of IPD, our base-case analysis indicated that the addition of this indirect effect yielded a substantially lower ICER (Table
3). The results of the current analysis consistently demonstrated a more favorable economic outcome when herd immunity was included in the vaccination decision model; however, it should be noted that the health-economic benefits of PCV-7 would be reduced by serotype replacement with uncovered virulent strains. Recent epidemiological data from the United States, Europe and Singapore suggest ongoing and significant serotype changes in children, with 19A in particular increasing in incidence [
61‐
63]. The adoption of PCVs with broader serotype coverage is strongly recommended. Meanwhile, continued surveillance of IPD is necessary to provide epidemiological data on potentially emerging serotypes.
The present analysis took a conservative approach to estimating the health benefits of PCV-7 vaccination for three reasons. First, the increasing frequency and rapid spread of antibiotic-resistant
S. pneumoniae is a global problem, and serious antibiotic resistance has been observed in Chinese clinical practice. According to an epidemiological study by Lin Zhou
et al., the rates of resistance to erythromycin and azithromycin in
S. pneumoniae in Beijing were 96.4 and 97.1%, respectively. Furthermore, 64.3% of all pneumococcal isolates were multidrug-resistant
S. pneumoniae (MDRSP) [
64]. To manage pneumococcal disease caused by antibiotic-resistant
S. pneumoniae, superior antibiotics and longer hospital stays are necessary, leading to increased consumption of health-associated financial resources. In the present analysis, resource savings from the prevention of antibiotic-resistant infection were not accounted for among the potential savings from PCVs because local data on the subject are inadequate and difficult to collect. The exclusion of such infections underestimates the cost-effectiveness of universal PCV-7 vaccination in China. Second, the most influential factor captured by our model was the
S. pneumoniae isolation rate for pneumonia. Our findings indicated that a higher isolation rate would yield more favorable health benefits from vaccination. The isolation rate used in the base-case analysis were 8%, which is greater than the estimates provided by Ying Chen
et al.; [
2] however, as a result of widespread antibiotic abuse in China, pathogen cultures from blood and cerebrospinal fluid (CSF) are rarely positive, and nasopharyngeal isolates are the main method of surveillance for pneumococcal epidemiology in Chinese children. According to a study by Rudan
et al., the estimated proportion of pneumococcal pneumonia in all-cause radiological and fatal pneumonia was 30–50% in Chinese children aged ≤5 years [
65]. One recent study conducted in Beijing found that nearly 55% of children with severe community-acquired pneumonia had
S. pneumoniae in their lung tissues, as identified by PCR and/or Southern blotting [
66]. Because of the limitations of these data, our model did not adjust for the isolation rate for
S. pneumoniae. These conservative estimates may have led us to underestimate the impact of a PCV program. Finally, several studies have indicated that the benefits of PCV7 vaccination extend beyond the covered serotypes. A separate study found a 20% reduction in non-pneumonia acute respiratory infections in children aged ≤2 years after the adoption of PCV-7 [
67]. Data from South Africa and the United States indicated that PCV may reduce the incidence of viral-associated pneumonia [
68]; however, our model does not consider these broader benefits of the vaccine and thus underestimates the impact of the vaccine.
This study has several weaknesses and limitations. First, in the absence of an epidemiological survey, no China-specific data were available for some of the model inputs, and there was substantial variation in data quality. Thus, such results must be carefully interpreted because multiple sensitive variables limit the extrapolation of results (e.g., the incidence of all-cause AOM,
S. pneumoniae isolation rate, and efficacy of herd immunity). Fortunately, our sensitivity analyses indicated that these variables do not significantly affect the ICERs in terms of the WHO-CHOICE standards. Second, to simplify the model structure and to avoid including too many uncertainties, the current model does not include all diseases associated with
S. pneumonia and adverse reactions related to the vaccination as other economic analyses do, such as sinusitis, septic arthritis and injection-side reactions [
21,
22,
69‐
71]. However, because the excluded diseases or events (e.g., seizure disorders and vision loss) are rare, their impact was too minor to be captured by the model. Third, almost all China-specific data in this analysis, such as the serotype coverage of PCV-7 for pneumonia, were obtained from hospital-based studies. These studies were mostly conducted in teaching hospitals in large cities such as Beijing, Shanghai and Guangzhou, where sanitary and health conditions are likely better than those in small cities and rural regions. Thus, such data may not be representative of the entire Chinese population, and there are uncertainties regarding the robustness of the analysis. If better supported and more representative epidemiological data become available in the future, the model outcome will be more accurate, and the relative effectiveness of the vaccines in preventing IPD will be greater. Fourth, potential uncertainties and biases result from the several assumptions and expert opinions used in this model, such as the efficacy of PCV-7 vaccination on a three-dose schedule and the productivity loss of parents who must care for their children. We adjusted these assumptions in the sensitivity analysis to determine the robustness of the model. Fifth, as other analyses have showed that herd immunity had a significant effect on cost-effectiveness, [
43] it should be carefully explained that the current analysis used the indirect efficacy data from other countries, where the patterns of mixing and serotype may differ greatly from that of China. Finally, changes in health-resource consumption and quality of life due to vaccine-related adverse events were not included in the current analysis, which may lead to an overestimation of the cost of disease. The ICER for vaccination would have been more unfavorable if such expenditures had been included; however, the available evidence from large-scale studies indicates that PCV-7 was well tolerated and that the impact on cost and quality of life would be minimal. Despite these shortcomings, our analysis provides insight into the potential of PCVs to reduce the burden of pneumococcal disease.
Competing interests
Declaration of personal interests: All authors have nothing to declare.
Authors’ contributions
BW contributed to the conception and design of the model and interpreted the results. DC developed the economic model, performed the analyses and drafted the manuscript. HZ and JH collected and reviewed data. All authors read and approved the final manuscript.