Avai lable onl ine at www.sc iencedirect .com

V A L U E I N H E A L T H R E G I O N A L I S S U E S 3 C ( 2 0 1 4 ) 1 5 6 – 1 6 6

2212-1099$36.00 &

Published by Elsevie

http://dx.doi.org/10.

Conflicts of inteNavarro-Locsin hascurrently an employthe Santa Ana Hosp

E-mail: sharon.x* Address correspo

journal homepage: www.elsevier .com/ locate /vhr i

Cost-Effectiveness Analysis of Pneumococcal Vaccination withthe Pneumococcal Polysaccharide NTHi Protein D ConjugateVaccine in the PhilippinesXu-Hao Zhang, PhD1,*, Maria Carmen Nievera, MD2, Josefina Carlos, MD3, Marilla Lucero, MD, PhD4,Gyneth Bibera, MD2, Maria Isabel Atienza, MD5, Oleksandr Topachevskyi, PGDip6,Cecilia Gretchen Navarro-Locsin, MD5

1GlaxoSmithKline Vaccines, Singapore; 2GlaxoSmithKline Vaccines, Makati City, Philippines; 3University of the East RamonMagsaysay Memorial Medical Centre, Quezon City, Philippines; 4Research Institute for Tropical Medicine, Muntinlupa City,Philippines; 5St. Luke’s Medical Centre, Quezon City, Philippines; 6GlaxoSmithKline Vaccines, Wavre, Belgium

A B S T R A C T

Objectives: To compare the cost-effectiveness of a universal massvaccination (UMV) program with a 2 þ 1 schedule of a 10-valentpneumococcal polysaccharide nontypeable Haemophilus influenzae pro-tein D conjugate vaccine (PHiD-CV) against two strategies: 1) a no-vaccination strategy and 2) a pneumococcal 13-valent conjugatevaccine (PCV13) 2 þ 1 strategy in the Philippines. Methods:A published Markov cohort model was adapted to simulate theepidemiological and economic burden of pneumococcal diseases(meningitis, bacteremia, pneumonia, and acute otitis media) withina projected birth cohort in 2012 of 1,812,137 newborns over lifetime.Analyses were conducted at an annual discount rate of 5% from theperspective of the Philippine government. The current evaluation wasupdated with the best available local/regional clinical epidemiologicaldata and published efficacy evidence. Results: Compared with the no-vaccination strategy, the PHiD-CV 2 þ 1 UMV program was projectedto prevent 3,343 deaths due to invasive pneumococcal diseases and

2014, International Society for Pharmacoeconomic

r Inc.

1016/j.vhri.2014.04.004

rest: X. Zhang, M.C. Nievera, and O. Topachevskyiserved as a speaker, an advisory board membeee of GlaxoSmithKline, but at the time of the studital. All the other authors declare that they have

.zhang@gsk.com.ndence to: Xu-Hao Zhang, GlaxoSmithKline Vaccin

Open access under CC BY-NC-ND license.

pneumonia and 326,862 cases of pneumococcal diseases, resulting inan incremental cost-effectiveness ratio of 50,913 pesos/quality-adjusted life-year gained, which was considered to be highly cost-effective according to the threshold recommended by the WorldHealth Organization. In comparison with the PCV13 2 þ 1 strategy,the PHiD-CV 2 þ 1 strategy was estimated to have a substantialreduction in acute otitis media (127,680 cases) and therefore a costsaving of potential 92.5 million pesos assuming price parity betweenPHiD-CV and PCV13 (US $1 ¼ 42.13 pesos in 2012). Conclusions: ThePHiD-CV 2 þ 1 UMV program is projected to be cost-effective,compared with no vaccination, and would provide substantial savingswith higher quality-adjusted life-year gains as compared with thePCV13 2 þ 1 strategy in the context of the Philippines.Keywords: cost-effectiveness, PCV13, PHiD-CV, Philippines, vaccination.

Copyright & 2014, International Society for Pharmacoeconomics andOutcomes Research (ISPOR). Published by Elsevier Inc.

Open access under CC BY-NC-ND license.

Background

The World Health Organization (WHO) estimated that 1.6 millionpeople die every year from invasive pneumococcal diseases(IPDs), and an estimated 0.7 to 1 million of these are childrenyounger than 5 years [1]. A study conducted in the Philippinesshowed that the peak incidence of IPDs is in children youngerthan 4 years. The same study estimated that one third of thechildren younger than 5 years with IPD would die [2].

Acute otitis media (AOM) is also a significant, although largelyunderestimated, public health burden among Filipino children.There is a paucity of good surveillance data regarding the extentof the burden of otitis media (OM) and/or AOM in Filipinochildren. A local study indicated that up to 25% of the children

with pneumonia had concomitant OM [3]. In a recent nationalsurvey, the overall prevalence of clinically diagnosed AOM inchildren aged 0 to 12 years in the Philippines was estimated at9.57%, with the 0 to 2 years age group having the most prevalentcases of AOM in the sample [4]. This prevalence rate is consideredas high on the basis of the WHO classification for the prevalenceof OM [5].

At present, two types of pneumococcal conjugate vaccines areavailable for use in children in the Philippines market. These areSynflorixTM (GlaxoSmithKline, Rixensart, Belgium), a pneumococcalpolysaccharide and nontypeable Haemophilus influenzae protein Dconjugate vaccine (PHiD-CV) and Prevnar 13TM (Wyeth LLC, Madi-son, NJ, now part of Pfizer, Inc.), a pneumococcal 13-valent con-jugate vaccine (PCV13). Both vaccines are licensed locally as a 2 þ 1

s and Outcomes Research (ISPOR).

are employees of the GlaxoSmithKline group of companies. C.G.r, and a primary investigator for GlaxoSmithKline. G. Bibera isy conduct and manuscript development, she was an employee ofno competing interests.es, 150 Beach Road, #22-00, Singapore 189720.

V A L U E I N H E A L T H R E G I O N A L I S S U E S 3 C ( 2 0 1 4 ) 1 5 6 – 1 6 6 157

scheme in the Philippines for the active immunization of infantsand children aged 6 weeks to 5 years against disease caused byStreptococcus pneumoniae (including meningitis, sepsis, bacteremia,pneumonia, and AOM) [6,7]. In addition, PHiD-CV is also licensed foruse against OM caused by nontypeable Haemophilus influenzae (NTHi)and licensed for use in preterm infants [6].

Health economics has been increasingly incorporated into thecomprehensive health technology assessment for formulary andvaccine policy decision making over the past decade. Publishedresults from other countries might not be directly transferablebecause of the sensitivity of key local data inputs such asepidemiology, local treatment patterns, or prices [8].

The purpose of this study was to evaluate the epidemiologicaland economic consequences of pneumococcal conjugate vacc-ine(s) for a universal mass vaccination (UMV) program in additionto the current standard of care for pneumococcal diseases inthe Philippines in 2012, using an adapted Markov cohort model.The two analyses presented in this article compared thecost-effectiveness of a UMV program with PHiD-CV 2 þ 1versus no vaccination, and that of PHiD-CV 2 þ 1 versus PCV13, inthe Philippines. The samemodel has been adapted for analyses in anumber of different countries such as Norway and Sweden [9,10].

Methods

Markov Cohort Model

A published Markov cohort model was adapted to simulate theepidemiological burden of pneumococcal and NTHi-related dis-eases, including IPDs, community-acquired pneumonia (CAP),and AOM, within a projected registered live birth cohort of1,812,137 newborns in the Philippines in 2012 [11,12]. Cohort-based analyses represent one of the most common forms of

Fig. 1 – Markov cohort model design. The cohort model is Marksequelae, and death. The transition from “no disease” to “sequelIn the model, only meningitis can lead to long-term sequelae; mbacteremia, respectively; and nonconsulting AOM are accountedmedia; GP, general practitioner; NTHi, nontypeable Haemophilus inflvalent conjugate vaccine; PHiD-CV, pneumococcal polysaccharide n

health economic modeling and are particularly useful for deter-mining the direct effect of medical interventions [13].

In the Markov cohort model, the individuals of the birthcohort move between the Markov states, as shown in Figure. 1,according to estimated transition probabilities derived from thepublished incidence rates, over a projection of lifetime horizon.The model has a number of mutually exclusive disease-relatedoutcomes, including meningitis, bacteremia, CAP, AOM, no pneu-mococcal infection, and death. Costs and quality-adjusted life-year (QALY) specific to each health state were estimated andsummarized over the cohort’s lifetime to calculate total accumu-lated costs and QALYs. The incremental cost and QALY gained ofthe two different strategies were computed as an incrementalcost-effectiveness ratio (ICER) for each set of comparison (PHiD-CVvs. no vaccination; PHiD-CV vs. PCV13). The WHO-recommendedthresholds for ICER were adopted in the study: 1) cost-saving: if thestrategy costs less and provides higher QALYs; 2) cost-effective:the ICER is below three times the gross domestic product (GDP) percapita of the country; 3) highly cost-effective: the ICER is below onetime the GDP per capita of the country; and 4) not cost-effective:the ICER is more than three times the GDP per capita of thecountry (GDP per capita of the Philippines in 2011 ¼ 103,366 pesos)[14]. It was assumed in the base-case scenario that 100% of thebirth cohort would be vaccinated according to the 2 þ 1 schedule(dose 1 at 6 weeks, dose 2 at 14 weeks, and dose 3 at 13 months).Finally, it was assumed that there was no herd effect for the UMVprogram of both vaccines in the base-case scenario.

Epidemiological Data

The birth cohort size of 2012 used in the analysis was 1,812,137,which was projected from the 2010 Field Health Service Informa-tion System report using a 1.9% population growth annually(National Statistics Office, Average population growth rate

ov-based with three exclusive health states: no disease,ae” or “death” is calculated on the basis of this decision tree.eningitis and bacteremia include NTHi meningitis and NTHifor in the quality-of-life impact calculation. AOM, acute otitisuenzae; PCP, primary care physician; PCV13, pneumococcal 13-ontypeable Haemophilus influenzae protein D conjugate vaccine.

Table 1 – Country-specific model parameters for the Philippines.

Parameter Philippines value

Size of birth cohort (2012 projected; cohortmodel)

1,812,137 [11]

Serotype distributionPneumococcal meningitis1. incidence rate o5 y (per 100,000) 18.3 [2]2. CFR o5 y (%) 30.8 [2]3. % cases with long-term sequelae o5 y

(%)11 [47]

Bacteremia1. Incidence rate o5 y (per 100,000) 14.06 (r1 y); 7.03 (2–4 y) [2]2. CFR o5 y (%) 33.3 [2]3. % cases with long-term sequelae o5 y

(%)10 [47]

All-cause pneumonia1. Hospitalization rate o5 y (per 100,000) 7957 (o1 y) (Lucero MG, unpublished ARIVAC data, 2012); 5003 (1 year) (Lucero MG,

unpublished ARIVAC data, 2012); 958 (2–4 y) [11,48,49]6.4 [21]

2. CFR o5 y (%) 8493 (o1 y) (Lucero MG, unpublished ARIVAC data, 2012); 4849 (1 year) (Lucero MG,unpublished ARIVAC data, 2012); 2424.5 (2–4 y) (extrapolation, Lucero MG, unpublishedARIVAC data, 2012)

3. GP consultation rate o5 y (per 100,000) 16412 (r1 y); 11745 (2–4 y) [4]AOM1. GP consultation rate o5 y (per 100,000) 98 (extrapolation based on expert opinion, 25% of the figure from the same age

group [20])2. Total number of myringotomy

procedures (per 100,000)See Appendix, Figure. 1

Discount rate (for the base-case scenario) 5%CFR for meningitis Age-specific data; see Appendix, Table 1% cases with long-term sequelae of

pneumococcal meningitisAge-specific data; see Appendix, Table 1

Incidence of bacteremia Age-specific data; see Appendix, Table 2CFR for bacteremia Age-specific data; see Appendix, Table 2% cases with long-term sequelae of

bacteremiaAge-specific data; see Appendix, Table 2

Hospitalization rate for all-causepneumonia

Age-specific data; see Appendix, Table 3

CFR for all-cause pneumonia Age-specific data; see Appendix, Table 3GP consultation rate for all-cause

pneumoniaAge-specific data; see Appendix, Table 3

GP consultation rate for AOM Age-specific data; see Appendix, Table 4Myringotomy procedure rate Age-specific data; see Appendix, Table 4AOM complication ratio (for sensitivity

analysis)Age-specific data; see Appendix, Table 4

Long-term sequelae rate for AOM Age-specific data; see Appendix, Table 4

Note. See Appendix in Supplemental Materials found at http://dx.doi.org/10.1016/j.vhri.2014.04.004.AOM, acute otitis media; ARIVAC, Acute Respiratory Infection Vaccine; CFR, case-fatality ratio; GP, general practitioner.

V A L U E I N H E A L T H R E G I O N A L I S S U E S 3 C ( 2 0 1 4 ) 1 5 6 – 1 6 6158

2000–2010, www.nscb.gov.ph) [11]. Age-specific annual incidencedata and case-fatality ratios for each pneumococcal/NTHi diseasewere obtained from the published literature or national data-bases from the pre-PCV7 period [2,15]. Table 1 and AppendixTables 1 to 4 (in Supplemental Materials found at http://dx.doi.org/10.1016/j.vhri.2014.04.004) present epidemiological data forIPD, bacteremia and sepsis, all-cause pneumonia, and AOM[2,4,15–21, Marcelo MJ. Case Series Report on Otogenic Complica-tions of Chronic Otitis Media. Philippine Children's MedicalCenter, unpublished data, 2004]. Local data were used whereveravailable. For certain parameters without local data, data fromneighboring country Malaysia with similar developing countrystatus, similar seasons, and similar epidemiology data wererevalidated by clinical experts and used for extrapolation of localdata if deemed suitable. For the long-term sequelae data of

invasive diseases, because of the lack of local or regional data,published data were used for extrapolations.

Vaccine Efficacy Parameters

The model assumed the vaccine efficacy to increase with theincreasing number of doses a child would receive, until maximalefficacy was achieved after the booster dose at month 13. Theduration of protection was assumed to be 10 years for the 2 þ 1regimen, with the linear waning of efficacy assumed to start atthe age of 2 years. Vaccine efficacy against IPD was calculated asa sum product using distribution of local serotypes (AppendixFigure. 1 in Supplemental Materials found at http://dx.doi.org/10.1016/j.vhri.2014.04.004) and serotype-specific efficacy of eachvaccine (Appendix Table 5 in Supplemental Materials found at

Table 2 – Vaccine-specific model parameters.

Parameter PCV13 value PHiD-CV value

Vaccine efficacy against IPD Serotype-specific efficacies were taken from aprevious study of PCV7 efficacy [22]

Serotype-specific efficacies were taken from aprevious study of PCV7 efficacy [22]

For serotypes not covered by PCV7, the meanefficacy of the PCV7 serotypes was used (94.7%).No efficacy was assumed for serotype 3 [23–29](Appendix, Table 5)

For serotypes not covered by PCV7, the meanefficacy of the PCV7 serotypes was used (94.7%).The reported rates of cross-protection forserotypes 6A (76%) and 19A (26%) were alsoincluded [23–29] (Appendix, Table 5)

Vaccine efficacy against all-cause pneumonia

Vaccine efficacy: reduction inGP visits (all-causepneumonia)

7.3% (assumed to be the same as PHiD-CV, due tothe lack of data)

7.3% [31]

Vaccine efficacy: reduction inhospitalized all-causepneumonia

23.4% (assumed to be the same as PhiD-CV, due tothe lack of data)

23.4% [31]

Vaccine efficacy against AOMVaccine efficacy: reduction in

AOM due to S. pneumoniaevaccine serotypes

57.2% (95% CI 44%–67%) [38] (assumed to be thesame as PCV7, due to the lack of data)

57.6% (95% CI 41.4%–69.3%) [37] (assumed to be thesame as an earlier 11-valent formulation,supported by data from well-established animalmodels)

Serotype replacement for S.pneumonia nonvaccineserotypes

�33.0% (�80.0%, 1.0%) [4,38] �33.0% (�80.0%, 1.0%) [38]

Vaccine efficacy: reduction inAOM caused by NTHi

0 35.3% (95% CI 1.8%–57.4%) [37]

Note. See Appendix in Supplemental Materials found at http://dx.doi.org/10.1016/j.vhri.2014.04.004.AOM, acute otitis media; CI, confidence interval; GP, general practitioner; IPD, invasive pneumococcal disease; PCV13, pneumococcal 13-valentconjugate vaccine; PHiD-CV, pneumococcal polysaccharide nontypeable Haemophilus influenzae protein D conjugate vaccine.

V A L U E I N H E A L T H R E G I O N A L I S S U E S 3 C ( 2 0 1 4 ) 1 5 6 – 1 6 6 159

http://dx.doi.org/10.1016/j.vhri.2014.04.004). To date, the mostrobust data on serotype-specific efficacy are the PCV7 resultsreported by Whitney et al. [22]. To the best of our knowledge, sofar there is no published meta-analysis comparing the efficacy ofPHiD-CV and PCV13. In addition, despite the fact that PHiD-CVhas been investigated in two large-scale randomized controlledtrials (Clinical Otitis Media & Pneumonia Study and FinnishInvasive Pneumococcal disease trial), no published randomizedcontrolled trial data are available for PCV13.

PHiD-CV and PCV13 were assumed to have the same serotypeefficacy as PCV7 for the seven serotypes covered by the latter. Forthe additional serotypes covered by PHiD-CV (i.e., serotypes 1, 5,and 7F) or PCV13 (i.e., serotypes 1, 3, 5, 6A, 7F, and 19A), theaverage vaccine efficacy of PCV7 vaccine serotypes (e.g., 94.7%)was used as the default value (Table 2). Cross-protection for 6A(76%) and 19A (26%) were assumed on the basis of PCV 7 vaccineeffectiveness data [22]. No efficacy was assumed for serotype 3contained in PCV13, as demonstrated by conflicting and hypo-responsive results on the effect of IPD in studies from differentcountries [23–28], which probably explains a statement made bythe UK Joint Committee on Vaccination and Immunisation that“PCV13 is effective against most of the additional six serotypes inPCV13 but not in PCV7. PCV13 appears ineffective against sero-type 3” [29].

In the absence of PCV13-specific pneumonia vaccine efficacyestimates, the Clinical Otitis Media & Pneumonia Study efficacyestimates of 23.4% for inpatient pneumonia and 7.3% for out-patient pneumonia for PHiD-CV were used for both PHiD-CV andPCV13 vaccines [30]. Despite several pneumonia efficacy trialswith vaccine formulations containing 7, 9, 10, and 11 serotypes[30–36], all the studies gave point efficacy estimates of between

20% and 35%, with no indication that vaccines with moreserotypes provided greater protection against pneumonia. Vac-cine efficacy against AOM was estimated on the basis of efficacyagainst pneumococcal vaccine serotypes and nonvaccine sero-type disease and efficacy against disease caused by NTHi. ForPHiD-CV, the assumptions were based on the data reported in thePneumococcal Otitis Efficacy Trial of an earlier 11-valent proteinD-conjugate vaccine formulation in prevention of AOM caused byboth S. pneumoniae and NTHi [37]. The assumptions of vaccineefficacy against AOM for PCV13 were based on an earlier random-ized double-blinded, clinical trial for PCV7 [38].

Using the Philippines epidemiology data and vaccine effec-tiveness data, a maximal effectiveness was calculated for eachvaccine using the following equation (Appendix Table 6 inSupplemental Materials found at http://dx.doi.org/10.1016/j.vhri.2014.04.004):

VEmax ¼ VEVT � % of AOM cases due to vaccine serotypes þVENVT � % of AOM cases due to nonvaccine serotypes þ VENTHi �% of AOM cases due to NTHi

The AOM complications ratio and the long-term sequelaerate were based on unpublished local data from tertiary hospi-tal. It is important to note that in the base-case scenario, wehave taken a more conservative stand to exclude the AOMcomplications and long-term sequelae from the analysis. Inthe sensitivity analysis, the AOM complications and long-termsequelae were included.

Health Outcomes

The current model was designed to estimate the overall impact ofdisease by considering QALY loss in both acute and long-term

Table 3 – Disutility weights of pneumococcal diseases.

Short-term Disutility weight Reference/assumptions

Meningitis (inpatient) 0.023 [39], value for meningitis with recoveryBacteremia (inpatient) 0.008 [39], value for hospitalizationPneumonia (inpatient) 0.008 Assumed to be the same as for inpatient bacteremiaPneumonia (outpatient) 0.006 [39], value for local infectionAOM (outpatient) 0.005 [40]AOM hospitalized myringotomy 0.005 Assumed to be the same as for AOM

Long-term Disutility weight Reference

Neurological sequelae from meningitis 0.400 [41]Hearing loss from meningitis 0.200 [41]Meningitis long-term sequelae (children) 0.269 [41]Meningitis long-term sequelae (adult) 0.286 [41]Bacteremia long-term sequelae (children) 0.269 Assumed to be the same as meningitisShort-term Disutility weight Reference/assumptions

AOM, acute otitis media.

V A L U E I N H E A L T H R E G I O N A L I S S U E S 3 C ( 2 0 1 4 ) 1 5 6 – 1 6 6160

sequelae. Because of the lack of relevant local data, disutilityweights, as shown in Table 3, which had been used in otherpublished PHiD-CV cost-effectiveness analyses, were used in thecurrent analyses [12,39,40].

Resource Use and Costs

The cost analyses were conducted from the perspective of thePhilippine government. Only direct medical costs (e.g., hospital-ization, inpatient/outpatient diagnostic tests and procedures,medication/vaccine costs, and health care professionals’ fees)were included. In the model, we have conservatively applied thesame price of 800 pesos (US $1 ¼ 42.13 pesos in 2012) per vaccinedose for both PHiD-CV and PCV13. The 800 pesos per vaccine doseprice was derived from the indicative tendering offer price ofGlaxoSmithKline for PHiD-CV in the Philippines. Cost data foracute episodes and related long-term sequelae of the IPD and

Table 4 – Costs for pneumococcal diseases.

Pneumococcal disease Averagecosts (PHP)*

Meningitis—first year (acute episode) 97,000 Expela

Meningitis long-term sequelae annual costfor subsequent years

23,500 Expela

Bacteremia—hospitalized (acute episode) 67,000 Expem

Bacteremia—outpatient (acute episode) 6,000 Expean

Bacteremia long-term sequelae annual costfor subsequent years

11,750(children)

Expem

18,170 (adult)Pneumonia—hospitalized (acute episode) 23,500 Expe

anPneumonia—outpatient (acute episode) 1,300 ExpeAOM hospitalized myringotomy 4,500 Expe

anAOM GP consultations (acute episode) 743 [42]

AOM, acute otitis media; GP, general practitioner; IV, intravenous; PHP,* US $1 ¼ 42.13 pesos.

CAP (Table 4) were based on expert validation of clinical treat-ment pathways and an average of the public and privatehospitals costs. AOM costs were based on locally published data[42]. Only direct medical costs were counted in the study. Healtheffects and costs were both discounted at 5% per annum.

Sensitivity Analyses

For both comparisons, extensive one-way sensitivity analyseswere performed to evaluate how robust the results are to thechange in model variables. These were performed using �20%(up to �50% depending on the inputs) for each of the base-casevalue of most of the variables; the upper and lower limits of the95% confidence interval were used for vaccine efficacy inputs. Wealso tested the effect of assuming the same steady herd effect of30% for the cohort on year 11 and beyond because we assumedthat the protection of both these vaccines lasts for 10 years. For

Reference

rts’ validated calculation, including room charge, medication, IV,boratory tests, and physician’s feerts’ validated calculation, including room charge, medication, IV,boratory tests, and physician’s feerts’ validated calculation, including room charge, laboratory tests,edication, and physician’s feesrts’ validated calculation, including consultation, laboratory tests,d medicationrts’ validated calculation, including room charge, laboratory tests,edication, and physician’s fees

rts’ validated calculation, including consultation, laboratory tests,d medicationrts’ opinions based on PhilHealth 2011 [49]rts’ validated calculation, including consultation fee, medication,d chest X-ray fees

Philippine pesos.

Table 5 – Estimated effect of PHiD-CV and PCV13 on lifetime disease burden (projected number of cases) in thePhilippines.

Healthoutcomes

No-vaccinationprogram (A)

PCV13 2 þ 1vaccinationprogram (B)

PHiD-CV 2 þ 1vaccinationprogram (C)

Difference PHiD-CV vs.no-vaccination program

(C � A)

DifferencePHiD-CV vs.PCV13 (C � B)

IPD cases (acuteepisode)

11,919 10,768 10,818 �1,102 50

All-causepneumonia cases(acute episode)

2,342,129 2,277,482 2,277,466 �64,664 �16

AOM cases (acuteepisode)

3,691,038 3,510,803 3,383,123 �261,096 �127,680

IPD cases (long-term sequelae)

3,378 3,299 3,302 �75 3

All deaths 39,574 1,773,057 1773,057 �3,343 0

AOM, acute otitis media; IPD, invasive pneumococcal disease; PCV13, pneumococcal 13-valent conjugate vaccine; PHiD-CV, pneumococcalpolysaccharide nontypeable Haemophilus influenzae protein D conjugate vaccine.

V A L U E I N H E A L T H R E G I O N A L I S S U E S 3 C ( 2 0 1 4 ) 1 5 6 – 1 6 6 161

the comparisons of PHiD-CV versus PCV13, three alternativescenario analyses were performed with

1.

T

C

VcA

L

T(uTcTQIncera

Ap*

different price of PCV13;

2. unpublished local AOM etiology with a high proportion of

NTHi-attributed AOM versus S. pneumonia–attributed AOM ascompared with the internationally published data; and

3.

inclusion of AOM long-term sequelae and complications.

We also conducted probabilistic sensitivity analysis to assessthe overall effect on results from simultaneous uncertaintiesaround input parameters, and present the results in a cost-effectiveness plane. Point estimates used in the base case as

able 6 – Direct costs by disease and vaccination progra

ost category No-vaccinationprogram (A)

PCV13 2 þ 1vaccinationprogram (B)

PHvp

accinationosts

0 4,331,835,868 4

cute episodeIPD 903,957,346 804,351,625 8All-cause

pneumonia13,432,362,973 12,304,532,946 12

AOM 3,145,091,640 3,031,573,906 2ong-termsequelaeIPD 2,835,267,365 2,643,669,639 2otalndiscounted)

20,316,679,324 23,115,963,985 23

otal discountedosts

10,273,482,597 13,219,953,924 13

otal discountedALYs gained

31,268,126 31,323,884

crementalost-ffectivenesstio

OM, acute otitis media; IPD, invasive pneumococcal disease; PCV13, polysaccharide nontypeable Haemophilus influenzae protein D conjugate vUS $1 ¼ 42.13 pesos.

input data were replaced by appropriate choice of distributions,for example, lognormal distribution for efficacy parameters,triangular distribution for cost inputs, and beta distribution fordisutility weights.

Results

Predicted Effect of Vaccination on Disease Burden and Costs

For the projected birth cohort of 1,812,137 in the Philippines in2012, it was estimated that without a UMV there would be a totalof 11,919 cases of IPD (3,513 cases of meningitis and 8,406 cases of

m in the Philippines (2012 Philippine pesos*).

iD-CV 2 þ 1accinationrogram (C)

Difference PHiD-CV vs.no-vaccination

program(C � A)

DifferencePHiD-CV vs.

PCV13(C � B)

,331,835,868 4,331,835,868 0

08,699,878 �95,257,468 4,238,253,304,464,588 �1,127,898,385 �68,358

,917,838,591 �227,253,049 �113,735,315

,652,113,618 �183,153,747 8,443,979,014,952,543 2,698,273,219 � 101,011,442

,127,492,075 2,854,009,477 �92,461,849

31,324,182 56,056 298

50,913 Dominant

neumococcal 13-valent conjugate vaccine; PHiD-CV, pneumococcalaccine; QALY, quality-adjusted life-year.

V A L U E I N H E A L T H R E G I O N A L I S S U E S 3 C ( 2 0 1 4 ) 1 5 6 – 1 6 6162

bacteremia), with 3,378 cases with long-term sequelae, 2,342,129cases of all-cause pneumonia, and 3,691,038 cases of AOM overtheir lifetime under the current standard of care for pneumo-coccal diseases (Appendix Table 7 in Supplemental Materialsfound at http://dx.doi.org/10.1016/j.vhri.2014.04.004). The associ-ated economic burden (discounted) amounted to be 10,273.48million pesos. For the base-case scenario, it was found that UMVwith PHiD-CV 2 þ 1 was highly cost-effective as compared withthe no- vaccination strategy for the 2012 birth cohort (n ¼1,812,137), with an ICER of 50,913 peso/QALY gained, which isless than one time the GDP per capita of the Philippines asrecommended by the WHO (equivalent to 103,366 pesos). Inanother comparison using the same cohort, it was projected thatcompared with PCV13, PHiD-CV 2 þ 1 would have a substantiallygreater reduction in the AOM cases (127,680 cases) and compa-rable reduction in IPD and CAP-related cases and IPD long-termsequelae (Table 5). Both vaccines would provide a broadly similareffect on reduction in the number of IPD-related and CAP-relateddeaths in the Philippines. The estimated financial projectionsbased on the same analysis showed that the PHiD-CV 2 þ 1vaccination program can reduce the financial burden of IPD, CAP,and AOM of the 2012 birth cohort (n ¼ 1,812,137) over lifetime(Table 6) by approximately 1.54 billion pesos against no vacci-nation. In comparison against the PCV13 vaccination program,the PHiD-CV vaccination program was estimated to result in totalsavings of approximately 92.5 million pesos while having anadditional gain of 298 QALYs, meaning 298 additional years in theperfect health state. Therefore, the PHiD-CV 2 þ 1 vaccinationprogram was evaluated to be a cost-saving and dominantstrategy as compared with the PCV13 2 þ 1 vaccination programfor the Philippines. This was indeed a conservative estimationwithout taking into account the complications and long-termsequelae associated with AOM. If the complications and

Fig. 2 – Tornado diagram for the top 10 most influential paramevaccination program versus no vaccination (one-way sensitivityconfidence interval; CAP, community-acquired pneumonia; GP,influenzae; PHiD-CV, pneumococcal polysaccharide nontypeable

long-term sequelae were incorporated into the base-case sce-nario, the total savings of PHiD-CV 2 þ 1 as compared with PCV13would be around 267 million pesos, with a higher QALY gainedof 364.

Sensitivity Analyses

The tornado diagram in Figure. 2 shows the effect of the 10 mostinfluential input parameters around the base-case value on thediscounted ICER (descending order), based on the extensive one-way sensitivity analyses. It was found that most of the modelvariables had very little influence over the conclusion that the PHiD-CV 2 þ 1 vaccination program as compared with the no-vaccinationstrategy was a highly cost-effective strategy for the Philippines.Among all the variables, only one variable (% reduction in CAPhospitalizations) had a greater effect on the final conclusion. Whenthe lower limit of the 95% confidence interval was selected, the ICERwas increased to 151,012 pesos/QALY gained, less than three timesthe GDP per capita of the Philippines (equivalent to 310,098 pesos).As expected, PHiD-CV 2 þ 1 became even more cost-effective thanthe no-vaccination strategy when a herd effect of 30% wasassumed, resulting in a lower ICER of 46,568 pesos/QALY.

The presentation of a one-way sensitivity analysis for adominant base case can be difficult to interpret using a tradi-tional tornado diagram because of the inability to calculate anICER. Therefore, Figure. 3 presents the effect of the nine mostinfluential variables on a cost-effectiveness plane generated bycomparing the PhiD-CV 2 þ 1 vaccination program with thePCV13 2 þ 1 vaccination program in the Philippines. It was foundthat most of the model variables had little influence over thecost-saving conclusion of the PhiD-CV 2 þ 1 vaccination program.As seen in Figure. 3, there are two variables/assumptions chang-ing the conclusion. These two observations, however, are not

ters of the cost-effectiveness analysis of PHiD-CV 2 þ 1analysis) in the Philippines. AOM, acute otitis media; CI,general practitioner; NTHi, nontypeable HaemophilusHaemophilus influenzae protein D conjugate vaccine.

Fig. 3 – Univariate sensitivity analysis comparing PHiD-CV with PCV13 in the Philippines. The effect of the nine mostinfluential variables on cost-effectiveness results of the PHiD-CV 2 þ 1 vaccination program versus the PCV13 2 þ 1 strategy.AOM, acute otitis media; CI, confidence interval; GP, general practitioner; LL, lower limit; NTHi, nontypeable Haemophilusinfluenzae; PCV13, pneumococcal 13-valent conjugate vaccine; PHiD-CV, pneumococcal polysaccharide nontypeableHaemophilus influenzae protein D conjugate vaccine; QALY, quality-adjusted life-year; UL, upper limit.

-₱10

₱0

₱10

₱20

₱30

₱40

-20 000 0 20 000 40 000 60 000 80 000 100 000 120 000

Cost

diff

eren

ce B

illio

ns

QALY Difference

Cost-effec�veness Plane

Individual simula�ons Base case Cost-effec�veness threshold

Fig. 4 – Cost-effectiveness plane of probabilistic sensitivityanalysis comparing PHiD-CV 2 þ 1 with the no-vaccinationstrategy in the Philippines. PHiD-CV, pneumococcalpolysaccharide nontypeable Haemophilus influenzae protein Dconjugate vaccine; QALY, quality-adjusted life-year.

V A L U E I N H E A L T H R E G I O N A L I S S U E S 3 C ( 2 0 1 4 ) 1 5 6 – 1 6 6 163

unexpected either. First, when the PhiD-CV efficacy was adjustedto the lower limit of the 95% confidence interval for PCV7for IPD, it was expected that fewer IPD cases could be prevented,resulting in a lower gain in QALY of the PhiD-CV strategyalthough the overall costs were still lower with PhiD-CV due tothe substantial savings by the prevention of AOM cases. Second,when the percentage of AOM cases due to S. pneumonia andNTHi was adjusted from 25.7% and 32.3% to 56% and 22%,respectively, the result showed less favorable QALYs of PhiD-CVwith lower cost savings compared with PCV13. This changewas in line with the key difference between the two vaccinesthat PhiD-CV provides better protection against AOM caused byNTHi. However, as anticipated, the herd effect of 30% assumedfor both vaccines did not alter the ICER result in the base-caseanalysis.

In addition, an alternative scenario analysis showed that ifthe price of PCV13 per dose was increased from 800 pesos to 900pesos, the discounted financial savings of PhiD-CV versus PCV13would be significantly increased to close to 620.34 million pesosbecause of the substantial increase in the vaccination costs.

Another scenario analysis was also performed with theunpublished local etiology data (2012) from Dr. GretchenNavarro-Locsin on AOM cultures collected from Philippine Gen-eral Hospital, Philippine Children’s Medical Centre, and St. Luke’sMedical Centre, with the NTHi-attributed AOM of 52.9% and S.pneumonia–attributed AOM of 23.5% (the base-case assumptionsbased on published data were as follows: NTHi-attributed AOM:32.3%; S. pneumonia–attributed AOM: 35.9%) [43]. The resultsshowed that as compared with PCV13, PhiD-CV would save162.1 million pesos, with incremental QALYs gained of 739.

The third scenario analysis showed that if the AOM long-termsequelae and complications were counted into the analysis, thetotal savings would be close to 267 million pesos, with incre-mental QALYs gained of 364.

-₱500

₱500

-1 500 -1 000 -500 0 500 1 000 1 500 2 000

Cost

diff

eren

ce M

illio

ns

QALY Difference

Cost-effec�veness Plane

Individual simula�ons Base case Cost-effec�veness threshold

Fig. 5 – Cost-effectiveness plane of probabilistic sensitivityanalysis comparing PHiD-CV 2 þ 1 with no PCV13 2 þ 1 inthe Philippines. PCV13, pneumococcal 13-valent conjugatevaccine; PHiD-CV, pneumococcal polysaccharidenontypeable Haemophilus influenzae protein D conjugatevaccine’ QALY, quality-adjusted life-year.

V A L U E I N H E A L T H R E G I O N A L I S S U E S 3 C ( 2 0 1 4 ) 1 5 6 – 1 6 6164

In the probabilistic sensitivity analysis, 99.9% of the simula-tions fell below the three times the GDP per capita threshold,indicating a high level of certainty that the PHiD-CV vaccinationprogram is cost-effective compared with the no-vaccinationstrategy (Fig. 4). In comparison against the PCV-13 vaccinationprogram, the PHiD-CV strategy was estimated to be very likely adominant strategy (74.2% of the simulations located in the south-east quadrant in Fig. 5: additional QALY gains with cost savings).

Discussion

Globally, pneumococcal and NTHi-associated diseases continueto result in substantial direct medical and caregiver costs despitecurrent PCV7 vaccination programs [43]. The evaluation showssubstantial health and economic burden of pneumococcal dis-eases on the health care system over the lifetime of the birthcohort without pediatric vaccination program against pneumo-coccal diseases. For the projected birth cohort of 1,812,137 new-borns in the Philippines in 2012, there would be a total of 11,919cases of IPD (3,513 cases of meningitis and 8,406 cases ofbacteremia), with 3,378 cases with long-term sequelae, 2,342,129cases of all-cause pneumonia, and 3,691,038 cases of AOM overtheir lifetime under the current standard of care for pneumo-coccal diseases without a UMV program in the Philippines. Theassociated economic burden (discounted) amounted to 10,273.48million pesos.

If the government were to adopt the PHiD-CV (2 þ 1 schedule)vaccination program as compared with the no-vaccination strat-egy, potentially 3,343 deaths due to IPD and CAP and 326,862cases of pneumococcal diseases (1,102 IPD cases, 64,664 pneumo-nia cases, and 261,096 AOM cases) would be prevented over thelifetime horizon of the study cohort. The cost-effectivenessanalysis also shows that the discounted ICER of the PHiD-CV 2þ 1 vaccination versus the no-vaccination program was 50,913peso/QALY gained, which is considered to be cost-effectivebecause it is less than one time the GDP per capita of thePhilippines.

When the PHiD-CV (2 þ 1) vaccination program is comparedwith the PCV13 (2 þ 1) strategy, the results show that an addi-tional 127,646 cases would be prevented with a reduction in AOM

cases due to NTHi AOM and a comparable reduction in IPD andCAP-related cases. This would further result in substantialfinancial savings of 92.5 million pesos with higher QALYs gained(298 additional years in perfect health state). Therefore, the studyshows that PHiD-CV is a cost-saving and dominant strategy ascompared with PCV13. It is also interesting to note that althoughthe incremental QALYs gained of 298 for the entire cohort overlifetime appear to be small, this improvement actually reflectssignificant disease burden reduction in AOM (n ¼ 127,680 cases)of PHiD-CV versus PCV13 because the disutility per episode weassumed in the data input for the AOM acute cases was quiteminimal (disutility of AOM ¼ 0.005).

Our findings are in line with those of the other published cost-effectiveness analysis to compare PHiD-CV vaccination programswith PCV13. By et al. [9] used a Markov cohort model to comparethe PHiD-CV 2 þ 1 strategy versus the PCV13 2 þ 1 strategy inSweden from a societal perspective. It was found that the PHiD-CV strategy would generate an additional 45.3 QALYs with asubstantial savings of estimated 62 million Swedish kroner (closeto US $9.3 million) for a cohort of 112,120 children [9]. Robberstadet al. [10] have also applied a Markov model to evaluate the cost-effectiveness of pneumococcal conjugate vaccines (PCV7, PCV13,and PHiD-CV) for a specific birth cohort (n ¼ 61,152) in Norway;PHiD-CV was also found to be a dominant strategy as comparedwith PCV13 with substantial savings of 24 million Norwegiankrone (close to US $4.15 million), with an additional 49 QALYsgained [10]. In the study published by Knerer et al. [12], it wasfound that PHiD-CV was also a dominant strategy as comparedwith PCV13, offering an additional savings of Can $9 million(close to US $9.2 million) in the birth cohort size of approximately33 million newborns in Canada, and additional savings of £4.9million (close to US $7.2 million) in the United Kingdom with thebirth cohort size of approximately 61 million newborns [15].

It is important to note that a conservative approach wasadopted in the base-case scenario analysis. It was assumed thatPHiD-CV possessed an efficacy against NTHi-derived AOM, butnot on IPD or CAP, the inclusion of which would result in higherfinancial savings with higher QALYs gained as compared withPCV13. The AOM-associated complications and long-term seque-lae were also not included in the base-case scenario. As men-tioned in the Results section, if AOM complications and long-term sequelae were included on the basis of local data reportedby Marcelo [Case Series Report on Otogenic Complications ofChronic OM. Philippines Children's Medical Center, unpublisheddata, 2004], the total savings of PHiD-CV compared with PCV13would be close to 267 million pesos, with a higher QALY gained of364 [9]. On a separate note, we believe that more total costsavings would be estimated if the analysis was conducted from asocietal perspective in which there is less productivity loss ofparents or caregivers for taking care of children suffering fromAOM for UMV with PHiD-CV. Similarly, the same trend ofanalyses would be observed if costs for other related expenses,for example, transport to and return from the hospitals andclinics, were included.

As is the case with all the modeling exercises, there are anumber of limitations to the current analysis as well. First,ideally, a dynamic model should be considered for the evalua-tions of vaccination against infectious diseases. In the absence ofreliable and published data required for constructing a dynamicmodel, however, an established Markov model for vaccinationagainst pneumococcal diseases was used. Given that recent datatend to suggest a similar herd effect between the two vaccines[44–46], we believe that the study analyses and findings wouldnot likely be much different from those reported in the articleshould a dynamic model be developed and used in this study.This assumption, however, may be changed with the availabilityof newer data. In this analysis, best locally published data were

V A L U E I N H E A L T H R E G I O N A L I S S U E S 3 C ( 2 0 1 4 ) 1 5 6 – 1 6 6 165

used wherever available, yet there were still some specific clinicalepidemiology variables that were not available locally. Similarly,local utility data, ideally, should be used in the analyses. Asalways, it is not available in many developing countries includingthe Philippines. So, published disutility weights have been con-sidered and used. Given that the relatively small values of thedisutility weights applied and also the relatively short episode ofeach of the infections included, we believe that the effect ofdisutility weights on QALY and hence cost-effectiveness ratios isnegligible. Furthermore, only a small proportion of patients wasassumed to have long-term disease sequelae and differencesbetween these proportions were small between the two vacci-nated groups. Published data or regionally available data setswere therefore reviewed by local clinical experts, and reasonableadjustments were made from these.

Conclusions

This economic modeling study showed that a PHiD-CV 2 þ 1universal vaccination program could potentially prevent a sub-stantial number of deaths and cases of pneumococcal diseases ascompared with the no-vaccination strategy. The results alsosuggested that the PHiD-CV 2 þ 1 vaccination program is likelyto be a cost-effective strategy and could potentially providesavings with higher health gain measured in QALYs as comparedwith the PCV13 2 þ 1 strategy. Monetary estimates based onMarkov cohort analyses could be used to assess the return oninvestment from vaccination programs of the national govern-ment and they could be considered in policy making.

Acknowledgments

We thank Emmanuelle Delgleize from GlaxoSmithKline Vaccinesfor critical review of the manuscript; Lakshmi Hariharan, JesseQuigley Jones, and Irene Chen from GlaxoSmithKline Vaccines formanuscript coordination; and Seng Chuen Tan from IMS HealthAsia for his help in preparing the manuscript.

Source of financial support: GlaxoSmithKline Vaccines fundedthis study and was involved in all stages of the study andanalysis. GlaxoSmithKline Vaccines also took in charge allcosts associated with the development and the publishing ofthe present article. All authors had full access to the data andretained full independence in preparing the article. Publication ofthe study results was not contingent on sponsor’s approval.

Supplemental Materials

Supplemental material accompanying this article can be found inthe online version as a hyperlink at http://dx.doi.org/10.1016/j.vhri.2014.04.004 or, if a hard copy of article, at www.valueinhealthjournal.com/issue(select volume, issue, and article).

R E F E R E N C E S

[1] Pneumococcal conjugate vaccine for childhood immunization: WHOposition paper [in English, French]. Wkly Epidemiol Rec 2007;82:93–104.

[2] Santos J, Madrid MAC, Banez MAP, et al. Active hospital-basedepidemiological surveillance of invasive pneumococcal disease (IPD) inchildren in Metro, Manila, Philippines. Presented at: 7th ISPPD. Tel Aviv,Israel, March 14-18, 2010.

[3] Lucero MG, Nohynek HM, Nillos LT, et al. AOM in Filipino children withpneumonia: a nested study within an efficacy trial of 11-valentpneumococcal conjugate vaccine in Bohol, Philippines. Presented at:7th ISPPD. Tel Aviv, Israel, March 14-18, 2010.

[4] Caro RM, Llanes EGDV, Ricalde RC. Prevalence of clinically diagnosedacute otitis media (AOM) in the Philippines: a national survey.Presented at: WSPID. Melbourne, Australia, November 16-19, 2011.

[5] Acuin J. Chronic Suppurative Otitis Media: Burden of Illness andManagement Options. Geneva, Switzerland: World HealthOrganization, 2004.

[6] GlaxoSmithKline. Summary of Product Characteristics: SynflorixSuspension for Injection in Pre-Filled Syringe. Uxbridge, Middlesex, UK:GlaxoSmithKline, 2012.

[7] Pfizer/Wyeth. Prevnar 13 full prescribing information. 2012. Availablefrom: http://labeling.pfizer.com/showlabeling.aspx?id=501. [AccessedMay 05, 2014].

[8] Transferability of economic evaluations across jurisdictions. Availablefrom: http://www.ispor.org/councils/documents/ISPOR_Report_Good_Practices_Econ_Data_Transferability0308.pdf.[Accessed May 05, 2014].

[9] By A, Sobocki P, Forsgren A, Silfverdal SA. Comparing healthoutcomes and costs of general vaccination with pneumococcalconjugate vaccines in Sweden: a Markov model. Clin Ther2012;34:177–89.

[10] Robberstad B, Frostad CR, Akselsen PE, et al. Economic evaluation ofsecond generation pneumococcal conjugate vaccines in Norway.Vaccine 2011;29:8564–74.

[11] FIield Health Service Information System. FIield Health ServiceInformation System (FHSIS) Annual Report 2010. Manila, Philippines:National Epidemiology Center, Department of Health, 2010.

[12] Knerer G, Ismaila A, Pearce D. Health and economic impact of PHiD-CVin Canada and the UK: a Markov modelling exercise. J Med Econ2012;15:61–76.

[13] Sonnenberg FA, Beck JR. Markov models in medical decision making: apractical guide. Med Decis Making 1993;13:322–38.

[14] Gross national income & gross domestic product. Available from:http://www.nscb.gov.ph/sna/2012/1st2012/2012per1.asp. [Accessed May5, 2014].

[15] Edmond K, Clark A, Korczak VS, et al. Global and regional risk ofdisabling sequelae from bacterial meningitis: a systematic review andmeta-analysis. Lancet Infect Dis 2010;10:317–28.

[16] Maimaiti N, Aljunid S, Ahmed Z, et al. Clinical burden of pneumococcaldiseases in Malaysia. Presented at: 14th Asia-Pacific Congress forPediatrics. Kuching, Malaysia, September 8-12, 2012.

[17] Yu VL, Chiou CC, Feldman C, et al. An international prospective studyof pneumococcal bacteremia: correlation with in vitro resistance,antibiotics administered, and clinical outcome. Clin Infect Dis2003;37:230–7.

[18] Feigin RD, Cherry JD. Feign and Cherry’s Textbook of PediatricInfectious Diseases. (6th ed.) Philadelphia, PA: Saunders, 2009.

[19] Centers for Disease Control and Prevention. Epidemiology andPrevention of Vaccine-Preventable Diseases. Public Health Foundation2012.

[20] Chang CJ, Wang PC, Huang YC, Wu BS. Economic and clinicalburden of pneumococcal diseases and acute otitis media inTaiwan: a nationwide population-based database analysis. Presentedat: ISPOR 4th Asia-Pacific Conference. Phuket, Thailand, September 5-7,2010.

[21] Lupisan SP, Ruutu P, Abucejo-Ladesma PE, et al. Central nervoussystem infection is an important cause of death in underfiveshospitalised with World Health Organization (WHO) defined severe andvery severe pneumonia. Vaccine 2007;25:2437244:

[22] Whitney CG, Pilishvili T, Farley MM, et al. Effectiveness of seven-valent pneumococcal conjugate vaccine against invasivepneumococcal disease: a matched case-control study. Lancet2006;368:1495–502.

[23] Butler JC, Breiman RF, Campbell JF, et al. Pneumococcal polysaccharidevaccine efficacy: an evaluation of current recommendations. JAMA1993;270:1826–31.

[24] Singleton RJ, Butler JC, Bulkow LR, et al. Invasive pneumococcal diseaseepidemiology and effectiveness of 23-valent pneumococcalpolysaccharide vaccine in Alaska native adults. Vaccine2007;25:2288–95.

[25] Miller E, Andrews NJ, Waight PA, et al. Effectiveness of the newserotypes in the 13-valent pneumococcal conjugate vaccine. Vaccine2011;29:9127–31.

[26] Andrews N, Miller E, Kaye P, et al. Effectiveness of the 13 valentpneumococcal conjugate vaccine against invasive pneumococcaldisease (IPD) in England and Wales. Presented at: 8th InternationalSymposium on Pnumococci and Pneumococcal Diseases. Iguacu Falls,Brazil. March 11-15, 2012.

[27] Dagan R, Patterson S, Juergens C, et al. Efficacy of 13-valent and 7-valent pneumococcal conjugate vaccines in preventingnasopharyngeal colonization with PCV7-serotypes: a randomizeddouble-blind trial. Presented at: 8th International Symposium onPnumococci and Pneumococcal Diseases. Iguacu Falls, Brazil, March11-15, 2012.

V A L U E I N H E A L T H R E G I O N A L I S S U E S 3 C ( 2 0 1 4 ) 1 5 6 – 1 6 6166

[28] Cohen R, Levy C, Bingen E, et al. Impact of 13-valent pneumococcalconjugate vaccine on pneumococcal nasopharyngeal carriage inchildren with acute otitis media. Ped Infect Dis J 2012;31:297–301.

[29] Pneumococcal sub-committee of Joint Committee on Vaccination andImmunisation: minute of meeting on Wednesday 30 May 2012.Available from: http://media.dh.gov.uk/network/261/files/2012/07/JCVI-minutes-Pneumococcal-sub-committee-meeting-held-on-30-May-2012.pdf. [Accessed May 05, 2014].

[30] Tregnaghi MW, Sáez-Llorens X, López P, et al. Evaluating the efficacy of10-valent pneumococcal non-typeable Haemophilus influenzae protein-Dconjugate vaccine (PHID-CV) against community-acquired pneumoniain Latin America. Presented at: 29th Annual Meeting of the EuropeanSociety for Paediatric Infectious Diseases. The Hague, The Netherlands,June 7-11, 2011.

[31] Cutts FT, Zaman SM, Enwere G, et al. Efficacy of nine-valentpneumococcal conjugate vaccine against pneumonia and invasivepneumococcal disease in The Gambia: randomised, double-blind,placebo-controlled trial. Lancet 2005;365:1139–46.

[32] COMPAS Group. Design/setting of COMPAS: Latin-American trialevaluating the efficacy of 10-valent pneumococcal non-typeableHaemophilus influenzae protein-D conjugate vaccine (PHiD-CV). Presentedat: 29th Annual Meeting of the European Society for Paediatric InfectiousDiseases. The Hague, The Netherlands, June 7-11 2011.

[33] Hansen J, Black S, Shinefield H, et al. Effectiveness of heptavalentpneumococcal conjugate vaccine in children younger than 5 years ofage for prevention of pneumonia: updated analysis using World HealthOrganization standardized interpretation of chest radiographs. PedInfect Dis J 2006;25:779–81.

[34] Klugman KP, Madhi SA, Huebner RE, et al. A trial of a 9-valentpneumococcal conjugate vaccine in children with and those withoutHIV infection. N Engl J Med 2003;349:1341–8.

[35] Lucero MG, Nohynek H, Williams G, et al. Efficacy of an 11-valentpneumococcal conjugate vaccine against radiologically confirmedpneumonia among children less than 2 years of age in the Philippines:a randomized, double-blind, placebo-controlled trial. Ped Infect Dis J2009;28:455–62.

[36] Madhi SA, Kuwanda L, Cutland C, Klugman KP. The impact of a 9-valent pneumococcal conjugate vaccine on the public health burden ofpneumonia in HIV-infected and -uninfected children. Clin Infect Dis2005;40:1511–8.

[37] Prymula R, Peeters P, Chrobok V, et al. Pneumococcal capsularpolysaccharides conjugated to protein D for prevention of acute otitismedia caused by both Streptococcus pneumoniae and non-typableHaemophilus influenzae: a randomised double-blind efficacy study.Lancet 2006;367:740–8.

[38] Eskola J, Ward J, Dagan R, et al. Combined vaccination of Haemophilusinfluenzae type b conjugate and diphtheria-tetanus-pertussis containingacellular pertussis. Lancet 1999;354:2063–8.

[39] Bennett JE, Sumner W II, Downs SM, Jaffe DM. Parents’ utilities foroutcomes of occult bacteremia. Arch Pediatr Adolesc Med 2000;154:43–8.

[40] Oh PI, Maerov P, Pritchard D, et al. A cost-utility analysis of second-lineantibiotics in the treatment of acute otitis media in children. Clin Ther1996;18:160–82.

[41] Morrow A, De Wals P, Petit G, et al. The burden of pneumococcaldisease in the Canadian population before routine use of the seven-valent pneumococcal conjugate vaccine. Can J Infect Dis Med Microbiol2007;18:121–7.

[42] Navarro-Locsin CG. Economic cost of OM in the Phil. Presented at: 6thExtraordinary International Symposium on Recent Advances in OtitisMedia. Seoul, South Korea, May 6-10, 2009.

[43] Talbird SE, Taylor TN, Caporale J, et al. Residual economic burden ofStreptococcus pneumoniae- and nontypeable Haemophilus influenzae-associated disease following vaccination with PCV-7: a multicountryanalysis. Vaccine 2010;28(Suppl. 6):G14–22.

[44] UK Pneumococcal Weekly Surveillance. Available from: http://www.Hpa.Org.Uk/Topics/InfectiousDiseases/InfectionsAZ/Pneumococcal/EpidemiologicalDataPneumococcal/CurrentEpidemiologyPneumococcal/.[Accessed May 5, 2014].

[45] Jokinen J, Rinta-Kokko H, Siira L, et al. Indirect impact of 10-valentpneumococcal conjugate vaccine against invasive pneumococcaldisease IPD among unvaccinated children in Finland. Presented at: 31stAnnual Meeting of European Society of Paediatric Infectious Diseases.Milan, Italy, May 28–June 1, 2013.

[46] Kilpi T, Palmu AA, Ruokokoski E, et al. Indirect impact of PneumococcalHaemophilus influenzae protein D conjugate vaccine (PHID-CV10) oninvasive pneumococcal disease (IPD) in a clinical trial. Presented at: 8thWorld Congress on Pediatric Infectious Diseases (WSPID). Cape Town,South Africa, November 19–22, 2013.

[47] Dagan R, Greenberg D, Jacobs MR, Phillips BL. Pneumococcal infections.In: Feigin RD, Cherry J, Demmler-Harrison GJ, Kaplan SL,eds. Feigin andCherry’s Textbook of Pediatric Infectious Diseases (6th ed.).Philadelphia: Elsevier Health Sciences, 2009.

[48] Republic of The Philippines. 2010 census of population and housinghighlights. Available from: http://www.census.gov.ph/statistics/census/population-and-housing. [Accessed May 5, 2014].

[49] Philippine Health Insurance Corporation: 2011 stats & charts. Availablefrom: http://www.philhealth.gov.ph/about_us/statsncharts/snc2011.pdf. [Accessed May 5, 2014].