Vaccine stockouts around the world: Are essential vaccines always
available when needed?
, Benjamin Schreiber
, Aurelia Gasca
, Laure Dumolard
, Daniela Urfer
, Kamel Senouci
World Health Organization, Expanded Programme on Immunization, Geneva, Switzerland
UNICEF Programme Division, New York, USA
UNICEF Supply Division, Copenhagen, Denmark
Introduction: As countries rise to the challenge of implementing the priorities of this ‘‘Decade of Vaccine”
and their commitments delineated in the Global Vaccine Action Plan (GVAP), many continue to face
important challenges of securing a continuous supply of essential vaccine for their national immuniza-
tion programme. This study provides evidence on the incidence of vaccine stockouts in countries, their
root causes and their potential impact on service delivery.
Methods: Vaccine stockout indicators collected from the WHO-UNICEF Joint Reporting Form (JRF) and
UNICEF’s Vaccine Forecasting Tool were analysed for the years covering the ﬁrst half of the GVAP
(2011 to 2015) and using 2010 as the baseline year. While the JRF collects annual information on national
and subnational stockouts by vaccine, the UNICEF Vaccine Forecasting Tool has the advantage of request-
ing UNICEF procuring countries to report on the reasons underpinning any stockouts.
Results: Every year on average, one in every three WHO Member States experiences at least one stockout
of at least one vaccine for at least one month. The incidence is most pronounced in Sub-Saharan Africa
where 38% of countries in this area of the world report national-level stockouts. The vaccines most
affected are DTP containing vaccines (often combined with HepB and Hib) and BCG. They account for
respectively 43% and 31% of stockout events reported. While national level vaccine stockouts occur in
countries of all income groups, middle income countries are the most affected. In 80% of cases, national
level stockouts were due to reasons internal to countries. More speciﬁcally, 39% of stockouts were attri-
butable to government funding delays, 23% were caused by delays in the procurement processes, and
poor forecasting and stock management at country level accounted for an additional 18%. When a
national level stockout of vaccines occurs, there is an 89% chance that a subnational stockout will occur
at district level. More concerning is that if a district level stockout occurs, this will lead to an interruption
of vaccination services in 96% of cases.
Discussion: There continues to be important challenges of ensuring a continuous availability of essential
vaccines. The global community, together with countries, urgently need to design effective interventions
aimed at reducing the frequency and mitigating the impact of stockouts.
Ó2017 Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creative-
One of the cornerstones of an effective national immunization
programme is for its supply chain to ensure a continuous and unin-
terrupted availability of essential vaccines up to the point of vacci-
nation. If vaccine availability is interrupted for any reason, missed
opportunities to vaccinate will occur and populations run the risk
of not being protected against deadly preventable diseases. Unfor-
tunately, recent evidence has shed light on the fact that vaccine
supply chain systems have gradually outgrown their ability to
ensure uninterrupted availability of vaccine and to manage this
current Decade of Vaccine priorities and commitments as delin-
eated in the Global Vaccine Action Plan (GVAP 2011–2020) [1–4].
As part of the GVAP monitoring and accountability framework, a
global vaccine availability indicator is being monitored yearly with
an overarching objective of a 2/3 reduction goals in the number of
countries with national level vaccine stockouts by 2020 [5,6].
In this study we present the ﬁndings from an in-depth analysis
of the GVAP vaccine stockout indicator in 194 WHO Member
States. The metric of vaccine availability used is the incidence
0264-410X/Ó2017 Published by Elsevier Ltd.
This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
Open Access provided for this article by the Gates Foundation.
E-mail address: firstname.lastname@example.org (P. Lydon).
Vaccine 35 (2017) 2121–2126
Contents lists available at ScienceDirect
journal homepage: www.elsevier.com/locate/vaccine
and duration of vaccine stockouts reviewed for the years covering
the ﬁrst half of the GVAP (2011–2015) together with a 2010 base-
line reference year. These indicators, while imperfect, are proxy
measures of a stressed vaccine supply chain and indicative of sys-
tems unable to ensure that essential vaccines are available when
needed. Disruptions in vaccine availability, places national immu-
nization programme at risk of not reaching their own targets and
goals of protecting their populations from vaccine preventable dis-
eases. The ﬁndings from this research will contribute to the limited
evidence on trends and performance of vaccine supply chain sys-
tems worldwide and serve as a ﬁrst step towards understanding
the root causes of vaccine stockouts. We hope that awareness of
the challenges will be raised in order for the global community,
together with countries, to design effective interventions aimed
at reducing the frequency and mitigating the impact of stockouts.
2.1. Data sources and years
The analysis was based on two main data sources. The ﬁrst is
the WHO-UNICEF Joint Reporting Form (JRF) . Since 2005 the
JRF collects data on the frequency of national level vaccine stock-
outs among a comprehensive set of systems indicators designed
to measure immunization performance and trends in member
state countries. Initially, three indicators were being reported by
WHO Member States as follows:
Was there any stockout of vaccine at national level? (Yes-No)
If ‘‘Yes”, please specify the duration of the stockout in months?
(duration in months or N/A)
Was there any stockout of vaccine at district level? (Yes-No)
In 2014, two indicators were added:
If ‘‘Yes” to the district level stockout question, was the district
level stockout caused by the national level stockout? (Yes-No
If district level stockouts occurred, were vaccination services
interrupted because vaccines were unavailable? (Yes-No
Countries respond to these ﬁve indicators
by vaccine as per
their national immunization schedule.
The second data source is the UNICEF Vaccine Forecasting Tool
. Countries that purchase vaccines through the UNICEF procure-
ment mechanism submit requested data in this tool as part of the
annual vaccine forecasting exercise managed by UNICEF Supply
Division. One of the reporting sections concerns the previous year’s
vaccine forecast and any occurrence of a stockout. While the initial
two indicators collected through the JRF mechanism are also
reported in the Vaccine Forecasting Tool,
the latter has the advan-
tage that it requests countries to report on the reasons for and causes
of the stockouts (Table 1).
For the ﬁrst two indicators related to stockouts occurrence at
national level and the causes of the stockouts, the periods covered
in this analysis are the baseline year of 2010 and the ﬁrst half of the
GVAP included between 2011 and 2015. The period of analysis for
the sub-national stockout indicators were the years 2014 and
2.2. Deﬁnitions and adjustments
In order to go deeper in the analytics than simply presenting
ﬁgures on the number of countries that report ‘‘Yes” to a vaccine
stockout, the notion of a stockout event becomes useful to deﬁne
when a stockout is reported for a speciﬁc vaccine. In other words,
if country reports ‘‘Yes” to experiencing a stockout for two separate
vaccines in the same reporting year, these are considered as two
separate stockout events. Likewise, if a country reports multiple
stockouts for the same vaccine in the same reporting year, each
is considered as a separate stockout event in the analysis and
quantiﬁed as such. Moreover, to be considered a full vaccine stock-
out, a minimum duration of 30 days (or 1 complete month) was
set. Given the global recommendation to maintain a 3 month buf-
fer stock of vaccines at national level at all times, a stockout would
indicate that these safety stocks have been depleted and vaccine
availability for the national immunization programme could be
compromised. If the reported duration of a stockout was less than
a month, this was not considered as a stockout event and was
excluded from the analysis.
The next adjustment was made in order to improve cross-
country comparisons and to improve reporting completeness. On
the former, an inclusion criteria was set for the analysis to focus
on a subset of vaccines that are common to all national immuniza-
tion schedules. Only stockouts reported for BCG, DTP, Measles con-
taining vaccines (ex: DTP-HepB-Hib or MMR), and Polio were
included in the analysis.
To improve reporting completeness and
ensure that the analysis can include all 194 WHO Member States, sev-
eral approaches were used to estimate missing information. The ﬁrst
method was to cross-check all indicators reported by each country in
a given year. In some instances, it was not uncommon for a country to
report on the duration of a stockout for a given vaccine but to forget
to respond ‘‘Yes” to the question of whether or not a stockout
occurred. This backward induction process was used to improve
the reporting completeness. The second method was to crosscheck
the vaccine product that is used in the country using the vaccine
schedule information from the JRF . The reason is that a country
may report a stockout for Hib vaccine, but not report it for HepB or
DTP. Yet the vaccine product used in their national immunization
programme is the pentavalent combination (DTP-HepB-Hib). As a
result, a retroﬁtting adjustment was performed so that other antigens
are included as part of the reported stockout event. The third method
was for WHO Regional Ofﬁce focal points to liaise back with the 25 to
30 countries that did not complete any information, and to solicit
responses to their JRF. For those that followed-up with the request,
it was to conﬁrm that no stockouts occurred in the reporting year.
Since most cases of a non-response by a country was to indicate that
no stockouts occurred, this was assumed for the remaining countries
that didn’t follow-up with WHO Regional Ofﬁce requests.
The reporting to UNICEF on the reasons for a national level
stockout needed were standardized and coded according to major
groups of root-causes. The reasons reported by countries were
grouped according to the following categories: (1) Financing
(delays in releasing funds to purchase vaccines), (2) Forecasting
and stock management (inaccurate forecast or poor stock manage-
ment), (3) Procurement (delays in the national procurement pro-
cess), (4) Product availability (shortages of vaccines on the global
market), (5) Quality (if a vaccine product was taken off the WHO
prequaliﬁcation list because of a quality concern and no longer
available), and (6) Other (all other causes speciﬁc to a country or
not identiﬁed/reported on).
The stockout indicator reference code for the JRF are Immunization System
Indicators [6280–6380 A-E].
The fact that both mechanisms requests the same information allows for
consistency checks to be made for the countries in common for both reporting
Newer vaccines that are not introduced in all countries (ex: Pneumococcal
conjugate or Rotavirus) or that are speciﬁc to certain regions of the world (ex: Yellow-
Fever or Japanese Encephalitis) were excluded in order to analyse countries across a
common set of vaccines used.
2122 P. Lydon et al./ Vaccine 35 (2017) 2121–2126
2.3. Analytical approach
The ﬁndings for this study cover three interrelated analytical
questions. This ﬁrst was to analyse how common vaccine stockout
events occur at national level and what vaccines are most affected.
The second was to understand the root-causes of these national
stockouts. The third was an attempt to shed light on the potential
impact of national level stockouts vaccine availability for service
delivery. The baseline year is set at 2010 – the year prior to the
start of the GVAP (2011–2020). The reference year for the analysis
is 2015 representing the mid-point to the 2020 GVAP goals (2011–
2015). The ﬁndings are presented in the aggregates, compared
across WHO regional groups, and looking at differences between
income groups. Figures are presented as simple averages.
3.1. National level stockouts around the world
At the mid-point in the GVAP, a total of 65 countries (34% of all
countries) reported at least one national-level stockout event for at
least one vaccine and for at least one month during 2015 (Table 2).
Since the 2010 baseline year, the situation improved very little
despite a promising downward trend during 2011–2014 and
where all progress was offset in 2015 – a year marking a setback
to reaching the GVAP target of a two-thirds reduction in the num-
ber of countries reporting national-level stockouts by 2020.
In countries reporting national-level stockouts, multiple events
often occur within the year when one or more vaccines are
affected. In some cases more than 4 stockout events are reported
by countries for a given year (Fig. 1). For the GVAP mid-point year
of 2015 a total of 113 stockout events were reported in the 65
countries - an average of 1.74 stockout events a year per country
reporting national level stockouts. The average stockout duration
was 44 days. The trend since 2013 indicates that the average dura-
tion of a stockout event is on the increase.
A review by groupings of countries uncovers the following addi-
During the ﬁrst half of the GVAP (2011–2015), 46% of countries
reported a maximum of 1 stockout event in any given reporting
year. For the mid-point year of 2015 this proportion reached
60% and indicates a slight trend away from multiple stockout
events reported by countries within a given year. An average
of 30% of countries reported more than 3 stockout a year with
a maximum reported in a country ranging between 6 and 7
stockout events in a single year.
Summary table of the key data sources.
Data source First/last year
No. of reporting
Stockout indicator scope Duration of
Reporting on causes
WHO-UNICEF Joint Reporting Form 2005–2015 Annual 194 National level by vaccine Months No
WHO-UNICEF Joint Reporting Form 2014–2015 Annual 194 Sub-national level by vaccine Months No
UNICEF Supply Division Vaccine
2002–2015 Annual 93 National level by vaccine Months Yes
National-level stockouts – Percent of countries by region, income and population.
2010 2011 2012 2013 2014 2015 Avg. 2011–15
No. countries reporting stockouts 67 66 57 54 50 65 58
% of countries 35% 34% 29% 28% 26% 34% 30%
No. of stockout events 153 148 120 112 110 113 121
Average duration of stockouts (days) 42 33 32 29 45 44 37
No. UNICEF procuring countries reporting stockouts 44 39 28 52 36 22 35
% of UNICEF procuring countries 49% 43% 31% 58% 40% 24% 39%
% countries by stockout frequency
One stockout event 43% 39% 40% 41% 50% 60% 46%
Two stockout events 19% 24% 25% 31% 18% 20% 24%
Three stockout events 13% 19% 24% 13% 12% 9% 15%
Four or more stockout events 25% 18% 11% 15% 20% 11% 15%
% stockout events by WHO Region
Americas (AMR = 35) 19% 18% 16% 17% 32% 17% 20%
Sub-Saharan Africa (AFR = 47) 39% 38% 38% 46% 32% 34% 38%
East Mediterranean (EMR = 21) 7% 11% 5% 6% 6% 9% 7%
Europe (EUR = 53) 18% 17% 18% 7% 14% 26% 16%
South-East Asia (SEA = 11) 4% 2% 5% 7% 4% 5% 5%
Western Paciﬁc (WPR = 27) 13% 14% 18% 17% 12% 9% 14%
% stockout events by Income
Low income (n = 31) 21% 23% 23% 26% 26% 20% 24%
Lower-Middle income (n = 50) 36% 29% 35% 41% 24% 28% 31%
Upper-Middle income (n = 58) 33% 32% 32% 29% 36% 27% 31%
High income (n = 55) 10% 16% 10% 4% 14% 25% 14%
% stockout events by Vaccine
BCG 33% 28% 34% 33% 25% 34% 32%
DTP containing vaccine 47% 45% 42% 35% 40% 51% 42%
Measles containing vaccine 7% 14% 9% 14% 14% 5% 11%
Polio 13% 13% 15% 18% 21% 10% 15%
Numbers in brackets indicate the total number of WHO Member States for each region respectively.
According to the World Bank classiﬁcation of countries. The middle income group includes countries that are classiﬁed as lower-middle income countries and upper-
middle income countries. Numbers in bracket indicate the number of WHO Member States in each income group.
P. Lydon et al. / Vaccine 35 (2017) 2121–2126 2123
The incidence of national-level stockouts is most pronounced in
Sub-Saharan Africa (AFR) where 38% of the 47 countries in this
region are affected. The next region most affected is the Amer-
icas. On the other hand, the South-East Asia and Western Paciﬁc
regions seem to be the least impacted by vaccine stockouts.
Stockouts at national level occur in countries of all income
groups. Both lower and upper middle income countries
reported the majority of stockouts – 31% on average for each
group between 2011 and 2015.
A deeper analysis by vaccine uncovers that during the ﬁrst half
of the GVAP, 43% of stockouts events concerned DTP-containing
vaccines – this proportion peaked in 2015 when 51% of all reported
stockouts events included DTP-containing vaccines (often com-
bined with HepB and Hib vaccine). The second vaccine most
affected by stockouts is BCG and represents another 31% of stock-
out events reported at national level.
3.2. Causes of national level stockouts
The analysis of the UNICEF Supply Division information uncover
that in 80% of cases, national level stockouts in the ﬁrst half of the
GVAP period were endogenous in nature. In other words, the rea-
sons were internal to the countries reporting stockouts rather than
due to exogenous, or external reasons outside of a country’s con-
trol. More speciﬁcally, 39% of stockouts were attributable to gov-
ernment funding delays (Fig. 2). Since UNICEF is unable to ship
vaccines to a country unless payment is received upfront, the com-
mon delays in government releasing funds to pay UNICEF for the
purchasing of vaccines invariably leads to a stockout. While vac-
cine ﬁnancing is the single largest root cause of stockouts, delays
in the procurement processes, poor forecasting and stock manage-
ment at country level accounted for a signiﬁcant share of internal
reasons for stockouts – respectively 23% and 18%.
For a remaining 11%, the cause of the stockout are exogenous
and independent to the country. In 9% of cases, the stockouts were
due to a global supply shortages. In rare cases, a quality issue on
certain vaccines caused countries to experience stockouts. This
would happen if a certain vaccine is no longer prequaliﬁed by
the WHO. In which case, UNICEF can no longer purchase this vac-
cine on behalf of countries and this can lead to a stockout.
3.3. Impact at subnational level and service delivery levels
In and of itself, a national level stockout does not necessarily
translate into a countrywide stockout of vaccines. There may be
instances where there is enough vaccines at subnational levels to
buffer against the effects of a national level stockout. In such case,
2010 2011 2012 2013 2014 2015
No. of countries reporng stockout events
Four or more stockouts
Fig. 1. Frequency of national level stockouts.
Poor forcasng &
Fig. 2. Root cause of national level vaccine stockouts in 90 UNICEF procuring countries (2011–15).
2124 P. Lydon et al. / Vaccine 35 (2017) 2121–2126
vaccination services are not interrupted by the lack of vaccines and
the national immunization programme is unaffected. That said, to
what extent do we know if this is the case? To shed light on the
potential impact of a national level stockout on subnational avail-
ability of vaccines and potential interruption in vaccination ser-
vices, a deeper analysis was conducted for the years 2014 and
The results indicate in 2015 a total of 58 countries reported
subnational-level stockouts at district level (Fig. 3). This represents
89% of the 65 countries that reported national level stockouts. Of
these 58 countries reporting subnational level stockouts, 86% (or
49 countries) indicated that the subnational level stockout was
caused by the national-level stockout. In other words, both
reported events were interlinked. For the remaining 14% of coun-
tries, the subnational-level stockout were caused by factors unre-
lated to the national-level stockout – a breakdown of the
distribution system or poor stock management at lower levels of
the supply chain for instance.
More concerning is that 47 of the 49 countries reported that the
district level stockout resulted in an interruption of vaccination
services. This implies that there is an 96% chance that a stockout
at district level will cause an interruption of immunization ser-
vices. In reviewing the trend over the two years of data, the situa-
tion deteriorated between 2014 and 2015 with more countries
reporting district-level stockouts and interruptions of vaccination
As the ﬁndings revealed, essential vaccines are not always avail-
able when needed and stockouts occur on a regular basis at both
national and subnational levels. Countries in all regions of the world
and of all income groups experience regular stockouts of key vacci-
nes, and in many cases these cause an interruptions of immuniza-
tion services. Results from an immunization coverage analysis in
17 countries that reported stockouts at all levels of the system
and interruptions in vaccination services suggest that stockouts
can potentially reduce DTP3 coverage by 7–10% points . In the
Philippines for instance, repeated stockouts of DTP-HepB-Hib vacci-
nes is estimated to have contributed to a 15% point reduction in the
national coverage estimate . Given the size of the population in
the Philippines, large segments became unprotected. It is no sur-
prise that Diphtheria cases were on the rise in 2015.
While this study has uncovered the ongoing challenges in
ensuring vaccine availability in countries, the interpretation of
these ﬁnding needs to be made with some understanding of the
key limitations. A ﬁrst set relate to the methodology that under-
pins the analysis and the need to make certain adjustments to
improve the reporting completeness or to improve cross-country
comparisons. Given that a national stockout of vaccine could be
an embarrassing situation for a country to report on, one would
expect that the incentive to under-report a stockout and/or its
duration would outweigh any reasons to over-report. Considering
further that some vaccines were excluded from the analysis, we
suspect that the ﬁndings presented in the analysis are likely to
underestimate the full magnitude of the problem at hand. This
underreporting potential was raised by the Strategic Advisory
Group of Experts (SAGE) for immunization in 2014 .
The second set of limitations relate to the indicators collected to
track subnational-level stockouts and whether these lead to an
interruption of vaccination services. While these indicators provide
valuable insights, the magnitude of the problem is difﬁcult to
gauge without a complete understanding of how many districts
were affected, or the extend of the interruption on vaccination ser-
vices and coverage performance. The third key limitation touches
on the root causes of stockouts – an analysis based on approxi-
mately 90 low and middle income countries that procure their vac-
cines through UNICEF and making the assumption that the root
cause results are generalizable. That said, it’s likely that exogenous
factors would play a greater role in explaining stockouts in middle
income countries that are self-procuring and often have difﬁculty
securing access to timely and affordable vaccines on the global
These limitations aside, this paper was an opportunity to assess
the situation half way through the GVAP and the trend is not mov-
ing towards the goal of a 2/3rd reduction in the number of coun-
tries experience at least one stockout of at least one vaccine for
at least one month. The target by 2020 is to have 25 countries or
less reporting national level vaccine stockouts. Yet, we remain far
from this target with 67 countries reporting stockouts in 2010
and 65 countries reporting stockouts in 2015.
Few would argue against the moral imperative to make vacci-
nes widely available on an equitable basis. This is especially true
for the poorest countries where the burden of vaccine preventable
diseases is the greatest. We see however, that many countries con-
tinue to face challenges of accessing an uninterrupted supply of
vaccines. Unless a concerted and coordinated effort is made to
address this problem, stockouts of life saving vaccines will con-
tinue to be commonplace. Effective interventions aimed at reduc-
ing the frequency and mitigating the impact of stockouts are
critically needed especially to address the root-causes of 75% of
national level stockouts namely: (1) addressing delays in releasing
national funds to purchase vaccines on time and ensure vaccine
ﬁnancing, (2) improving forecasting accuracy, stock management
practices and data systems for managing vaccines, (3) addressing
cumbersome national procurement processes and delays, espe-
cially in self-procuring countries. Recent discussion at global level
on these issues hold the promise for a change [11,12]. In the mean-
time, we hope that this analysis serves to raise awareness on the
vaccine stockouts situation worldwide and initiate interventions
aimed at reducing and eliminating causes of future ones.
Conﬂict of interest
All authors declare no conﬂict of interest.
 Humphreys G. Vaccination: rattling the supply chain [news]. Bull World
Health Organ 2011;89:324–5. http://dx.doi.org/10.2471/BLT.11.030511.
stockouts due to
Total Number of Countries
Fig. 3. Sub-national impact of national level vaccine stockouts in 2014 and 2015.
P. Lydon et al. / Vaccine 35 (2017) 2121–2126 2125
 Zaffran M, Vandelaer J, Kristensen D, Melgaard B, Yadav P, Antwi-Agyei KO,
et al. The imperative for stronger vaccine supply and logistics systems.
Vaccine 2013;31(Supplement 2):B73–80. http://dx.doi.org/10.1016/j.vaccine.
 World Health Organization (WHO). Immunization Supply Chain and Logistics – A
neglected but essential system for national immunization programmes. A Call-to-
Action for National Programmes and the Global Community. WHO
Immunization Practices Advisory Committee (IPAC). WHO/IVB/14.05 <http://
[accessed 31 August 2016].
 World Health Organization [Internet]. Global Vaccine Action Plan 2011–2020.
Geneva: WHO (2013). Available from: <http://www.who.int/immunization/
global_vaccine_action_plan/GVAP_doc_2011_2020/en/> [accessed 31 August
 World Health Organization [Internet]. 2015 Assessment Report of the
Global Vaccine Action Plan. Strategic Advisory Group of Experts on
Immunization (SAGE) <http://www.who.int/immunization/global_vaccine_
31 August 2016].
 World Health Organization [Internet]. 2014 Assessment Report of the Global
Vaccine Action Plan. Strategic Advisory Group of Experts on Immunization
31 August 2016].
 World Health Organization [Internet]. <http://www.who.int/
WHO_UNICEF_JRF_EN.xls> [accessed 31 August 2016].
 UNICEF Supply Division. <http://www.unicef.org/supply/index_55506.html>
[accessed 31 August 2016].
 World Health Organization (WHO). Unpublished internal report for the SAGE
working group on GVAP.
 WHO/UNICEF estimates of national immunization coverage (WUENIC) for the
Philippines. Accessible at: <http://apps.who.int/immunization_monitoring/
globalsummary/JPG/PHLDTP3_WUENIC.jpg> [accessed 31 August 2016].
 World Health Organization (WHO). Pre-empting and responding to vaccine
supply shortages. Strategic Advisory Group of Experts (SAGE). April 2016.
Accessible at: <http://www.who.int/immunization/sage/meetings/2016/april/
presentations_background_docs/en/index1.html> [accessed 31 August 2016].
 World Health Organization (WHO). World Health Assembly. Addressing the
global shortage of medicines and vaccines. WHA Resolution A69.25. May 2016.
Accessible at: <http://apps.who.int/medicinedocs/en/d/Js22423en/> [accessed
31 August 2016].
2126 P. Lydon et al. / Vaccine 35 (2017) 2121–2126