Bull World Health Organ 2021;99:783–794D | doi: http://dx.doi.org/10.2471/BLT.20.285258
Research
783
IntroductionRwanda has a strong immunization system,1 a well-organized vaccine supply system and a well-functioning community health worker (CHW) programme.2 As a result, national im-munization coverage rates are high and new vaccines are introduced swiftly.3,4 Nevertheless, some communities still face local disease outbreaks due to underimmunization. A major driver of underimmunization is vaccine hesitancy,5 which is defined by the World Health Organization (WHO) as “delay in acceptance or refusal of vaccination despite availability of vaccination services.”6 According to WHO, “vaccine hesitancy is complex and context specific, varying across time, place, and vaccines” and is one of the top 10 global health threats.6,7
Factors influencing vaccine hesitancy can be categorized using the 3Cs framework: (i) confidence; (ii) complacency; and (iii) convenience.8 As psychological, sociological and environ-mental drivers are paramount in instigating the behavioural changes required to address vaccine hesitancy,9,10 understand-ing the contextual relationships between vaccine hesitancy and socioeconomic determinants of health is crucial.11–16 Additionally, better understanding of vaccine hesitancy is needed to achieve the sustainable development goals, including universal access to quality vaccines.17 WHO’s Immunization Agenda 2030 highlights the importance of people-centredness for understanding the context-specific root causes of vaccine hesitancy and for co-designing solutions.18,19 However, the literature has major gaps on: (i) the analysis of human-centred design approaches; (ii) the relationship between beneficiaries
and service delivery; and (iii) incorporating the interaction between hesitancy factors into policy and intervention design.6
For our study of underimmunization drivers in a low-income country, we focused on measles in Rwanda, because: (i) measles is highly contagious and there were indications of underimmunization; (ii) measles increases morbidity and mortality because it erases the immune memory and increases susceptibility to other infectious diseases;20,21 (iii) despite long-standing vaccination programmes, eradication has not been achieved globally according to the Global Measles and Rubella Strategic Plan 2012–2020;22 and (iv) the second measles vaccine dose was designated a performance tracer in the Immunization Agenda 2030.18 Moreover, the measles programme is monitored by WHO and the Rwandan govern-ment, which can provide data for future quantitative models. In Rwanda, the measles vaccination programme has been quite successful: it achieved a national coverage rate above 85% for the second measles and rubella vaccine dose in 2015 and reached the target national coverage rate of 95% for the first dose in 2017.18,23 The first measles and rubella vaccine dose is administered 9 months after birth and, since 2015, the second dose is administered 15 months after birth. Despite these achievements, Rwanda still faces sporadic measles outbreaks (e.g. in Western Province in 2019).24
To inform policy design and to improve immunization levels, we aimed to study the mechanisms underlying vac-cine hesitancy in Rwanda that contributed to local underim-munization for measles and a subsequent measles outbreak. We conducted an analysis of immunization service delivery
Objective To investigate vaccine hesitancy leading to underimmunization and a measles outbreak in Rwanda and to develop a conceptual, community-level model of behavioural factors.Methods Local immunization systems in two Rwandan communities (one recently experienced a measles outbreak) were explored using systems thinking, human-centred design and behavioural frameworks. Data were collected between 2018 and 2020 from: discussions with 11 vaccination service providers (i.e. hospital and health centre staff ); interviews with 161 children’s caregivers at health centres; and nine validation interviews with health centre staff. Factors influencing vaccine hesitancy were categorized using the 3Cs framework: confidence, complacency and convenience. A conceptual model of vaccine hesitancy mechanisms with feedback loops was developed.Findings A comparison of service providers’ and caregivers’ perspectives in both rural and peri-urban settings showed that similar factors strengthened vaccine uptake: (i) high trust in vaccines and service providers based on personal relationships with health centre staff; (ii) the connecting role of community health workers; and (iii) a strong sense of community. Factors identified as increasing vaccine hesitancy (e.g. service accessibility and inadequate follow-up) differed between service providers and caregivers and between settings. The conceptual model could be used to explain drivers of the recent measles outbreak and to guide interventions designed to increase vaccine uptake.Conclusion The application of behavioural frameworks and systems thinking revealed vaccine hesitancy mechanisms in Rwandan communities that demonstrate the interrelationship between immunization services and caregivers’ vaccination behaviour. Confidence-building social structures and context-dependent challenges that affect vaccine uptake were also identified.
a Research Center for Access-to-Medicines, Naamsestraat 69, 3000 Leuven, Belgium.b East African Community Regional Centre of Excellence for Vaccines, Immunization and Health Supply Chain Management, University of Rwanda, Kigali, Rwanda.c System Dynamics Group, University of Bergen, Bergen, Norway.d Leuven University Vaccinology Center, KU Leuven, Leuven, Belgium.e Massachusetts General Hospital Institute for Technology Assessment, Harvard Medical School, Cambridge, United States of America.Correspondence to Catherine Decouttere (email: catherine.decouttere@ kuleuven .be).(Submitted: 31 December 2020 – Revised version received: 5 August 2021 – Accepted: 5 August 2021 – Published online: 28 September 2021 )
Insights into vaccine hesitancy from systems thinking, RwandaCatherine Decouttere,a Stany Banzimana,b Pål Davidsen,c Carla Van Riet,a Corinne Vandermeulen,d Elizabeth Mason,e Mohammad S Jalalie & Nico Vandaelea
784 Bull World Health Organ 2021;99:783–794D| doi: http://dx.doi.org/10.2471/BLT.20.285258
ResearchVaccine hesitancy mechanisms, Rwanda Catherine Decouttere et al.
in both a rural community and a peri-urban community between 2018 and 2020 and derived a conceptual model of vaccine hesitancy to assist in the design of sustainable interventions.
MethodsOur study design was based on WHO’s framework for health system building blocks and an already published im-munization system diagram (Fig. 1).25,26 In analysing vaccine uptake, we applied systems thinking and behavioural frameworks such as the 3Cs framework and the behavioural drivers theory.8,9,26–31 First, we assessed how underimmu-nization was influenced by the three so-called immunization service flows: (i) the vaccinee (child); (ii) the health-care workforce (nurse); and (iii) vaccine availability (vaccine). Second, we con-ducted in-depth interviews with health centre staff and collected secondary data (e.g. on vaccine orders, invento-ries and disease outbreaks). Based on our findings, we interviewed children’s
caregivers to understand local factors influencing vaccine hesitancy. Then, we used the 3Cs framework to categorize factors reported by caregivers and health centre staff.6 Finally, we derived causal relationships between behavioural drivers, vaccination intent and vacci-nation uptake by analysing the vaccine hesitancy factors identified and present these relationships in causal loop dia-grams. These diagrams are helpful for understanding complex systems and for developing interventions.29,30 All causal relationships in the causal loop diagrams and vaccine hesitancy factors were vali-dated and explained during additional discussions with health centre staff.
Participants
We interviewed 11 vaccinators and staff of the Expanded Programme on Immu-nization at three district hospitals, six health centres in a peri-urban setting in Kicukiro District, Kigali Province and five health centres in a rural setting in Ngororero District, Western Province. Additional details are available from
the data repository.32 At the community level, we interviewed caregivers at two health centres: one in rural Ramba in Ngororero District, where there was a measles outbreak in 2019, and one in Gahanga in the Kicukiro District. After a measles outbreak in July 2019, one of the Ramba health centre’s outreach posts became the Sovu health centre to be closer to people in the catchment area. Although these two locations share the same climate, epidemiologi-cal characteristics and health systems, they have different geographical and socioeconomic characteristics (Table 1). We attended two vaccination sessions at health centres and one outreach vac-cination session in both communities. Each caregiver present was invited to participate in the study (i.e. a conve-nience sample) by having a face-to-face interview in or near the session waiting room. After the purpose of the research was explained, 161 caregivers partici-pated: they were predominantly female and included both experienced mothers and mothers of firstborns.
Fig. 1. Immunization system diagram
R1
Population health status
Vaccine distribution
–
+ Outbreak response immunization
R: Reinforcing loop B: Balancing loop
R3
R2
R4
B2
B1
+
+
+
+
+
+
+
++ +
++
+
Vaccine supply
VaccinationImmunization
plan
Population well-being
Planned immunization Emergency immunization
Workforce and infrastructure
Surveillance and response
Population at risk
Morbidity and mortality
Outbreak
–=
=
=
==
+
+ Positive effect – Negative effect = Time delay in effect
+
+
Notes: The immunization system diagram is a high-level conceptual model of a national immunization system that was developed to engage stakeholders in immunization system assessment and sustainable redesign.25 The left side illustrates planned immunization (i.e. routine immunization and planned immunization campaigns) guided by a national immunization plan, and the right side illustrates emergency immunization (i.e. immunization in response to disease outbreaks). Loops R1 to R4 are reinforcing feedback loops that cause change in the same direction and loops B1 and B2 are balancing loops that cause change in the opposite direction. See Decouttere et al. for more information.25 The focal point of the immunization system diagram is vaccination and all efforts in the system are geared towards service delivery. Vaccination depends on three fundamental so-called flows that need to be synchronized at every vaccination session: (i) child (represented in the diagram by population well-being); (ii) nurse (represented by workforce and infrastructure); and (iii) vaccine (represented by vaccine distribution). Successful vaccination leads to better population health status, although only after a delay. From population health status, there are several loops that affect both planned and emergency immunization, but in different ways.
785Bull World Health Organ 2021;99:783–794D| doi: http://dx.doi.org/10.2471/BLT.20.285258
ResearchVaccine hesitancy mechanisms, RwandaCatherine Decouttere et al.
Data collection and analysis
In-depth discussions were conducted with staff at the 11 health centres (data repository)32 between September 2018 and November 2019 by three senior re-searchers and between December 2019 and February 2020 by one researcher. Discussions lasted 60 to 120 minutes and followed a topic guide (more details available in the data repository).32 Four researchers analysed data in field notes and photographed documents.
After one day’s training, six Rwan-dan data collectors conducted semis-tructured interviews with caregivers in November 2020 (data repository).32 Conversations lasting 7 to 12 minutes were held in Kinyarwanda, translated into English by the interviewer and later discussed with a supervisor and the research team to reach a consensus on interpretation. Then, two researchers analysed factors associated with vac-cine hesitancy using Excel (Microsoft Corporation, Redmond, United States of America) and four researchers clas-sified these factors. One researcher held additional discussions with health centre staff between April and Decem-ber 2020 to validate the findings (data repository).32
Three researchers used the main underlying mechanisms of vaccine hesitancy identified by the analysis to construct causal loop diagrams, which illustrate the interconnections between different factors and show feedback loops. The feedback loops can be ei-ther reinforcing (i.e. cause change in the same direction) or balancing (i.e. cause change in the opposite direction). The study was approved by the Rwanda National Ethics Committee (No. 195/RNEC/2019) and is reported according to consolidated criteria for reporting qualitative research.34
ResultsDuring data collection at the 11 health centres, we found no evidence of vac-cine stock-outs in 2018 or 2019. More-over, there were no human capacity limitations that resulted in immuniza-tion services being unavailable and national immunization coverage rates were above 90%. However, measles outbreaks occurred because of a lack of timely immunization. The suspected cause was reluctant vaccine uptake rather than limited vaccine availability.
Vaccine hesitancy
Service providers’ perspective
Interviews with Expanded Programme on Immunization staff at the Ramba and Sovu health centres, the Gahanga health centre and nine neighbouring health centres revealed contextual and behavioural factors that contributed to vaccine hesitancy. These factors were categorized according to the 3Cs frame-work (i.e. confidence, complacency and convenience) using a determinants matrix (Table 2; available at https:// www .who .int/ publications/ journals/ bulletin/ ).6 Vaccination service providers reported that confidence is increased by trust in immunization service delivery and that a good relationship between caregiv-ers and both nurses and CHWs is key. However, the time health centre staff could spend with each caregiver and the resulting quality of care were affected by an increased workload due to high peri-urban immigration rates of up to 10% per year, paperwork and other responsibilities.
We found that complacency was successfully reduced by: (i) organizing 6-monthly mother-and-child health weeks; (ii) regular community meet-ings (umuganda and w'ababyeyi or mothers’ evenings); (iii) educational sessions before vaccination sessions; and (iv) individual contacts between CHWs and caregivers. However, in rural settings, mothers tended to deprioritize or forget immunization of older children because of other tasks and because they had little contact with health centres in the 6-month interval between the first and second measles and rubella vaccine doses. This factor increased compla-cency and, combined with low disease surveillance in Sovu (which led to un-detected measles cases), contributed to the 2019 measles outbreak. CHWs were regarded as playing a crucial role in con-necting mothers to antenatal care and vaccination services. However, concerns were raised about the sustainability of CHW programmes.
With regard to convenience, the availability of immunization services
Table 1. Characteristics of two participating communities, study of vaccine hesitancy, Rwanda, 2018–2020
Characteristic Community
Ramba and Sovua Gahanga
Geographical context Rural community, hilly and remote landscape, limited road infrastructure and long distances to the health centre (i.e. more than 5 km or 2 hours travel time)
Peri-urban community with relatively short distances to the health centre (i.e. less than 5 km or 2 hours travel time)
Socioeconomic context
22% of the population live in extreme poverty,33 and many men are employed in coltan mining
9% of the population live in extreme poverty,33 and the affordability of transport is less problematic than in Ramba or Sovu
Location of local district hospital
Kabaya, Ngororero District, Western Province
Masaka, Kicukiro District, city of Kigali
Population 64 000 (50% Ramba health centre and 50% Sovu health centre)
67 000
Weekly vaccination sessions at fixed locations or outreach posts
2–4 at the Ramba health centre and 1–3 at the Sovu health centre (since 1 January 2020)
1–2 (fewer outreach services due to shorter distances and more affordable transport)
Measles outbreak since 2018
July 2019 in the Sovu catchment area
None
Migration No significant in or out migration
Increasing number of people settling in the area as they flee Kigali’s city centre where property prices are continuously increasing. Some clients visited health centres other than the one they were assigned to
a After a measles outbreak in July 2019 at one of the Ramba health centre’s outreach posts, that post became the Sovu health centre to be closer to people in the catchment area.
786 Bull World Health Organ 2021;99:783–794D| doi: http://dx.doi.org/10.2471/BLT.20.285258
ResearchVaccine hesitancy mechanisms, Rwanda Catherine Decouttere et al.
Tabl
e 3.
Ca
regi
vers
’ com
men
ts o
n fa
ctor
s affe
ctin
g va
ccin
e he
sitan
cy, R
wan
da, 2
018–
2020
Fact
or a
ffect
ing
vacc
ine
hesit
ancy
a
Sele
cted
com
men
ts o
n fa
ctor
by c
areg
iver
sb,c,d
,e
Rura
l hea
lth ce
ntre
s in
Ram
ba a
nd Su
vof (n
= 74
)Pe
ri-ur
ban
heal
th ce
ntre
in G
ahan
gag (n
= 87
)
Confi
denc
e
Tr
ust i
n th
e eff
ectiv
enes
s and
sa
fety
of v
acci
nes a
nd in
thei
r m
anuf
actu
rers
Posit
ive:
(i) a
ll re
spon
dent
s wer
e ha
ppy
to v
acci
nate
thei
r chi
ld a
s it p
rote
cts t
he c
hild
aga
inst
dise
ase
(of
cour
se, s
elec
tion
bias
mus
t be
cons
ider
ed h
ere
as c
areg
iver
s at v
acci
natio
n se
ssio
ns w
ere
inte
rvie
wed
. H
owev
er, r
espo
nden
ts m
entio
ned
that
, “not
hing
cou
ld st
op m
e to
com
e to
the
serv
ice”
or c
alle
d th
e se
rvic
es “a
ble
ssin
g,” w
hich
show
ed st
rong
mot
ivat
ion
and
ampl
e co
nfide
nce)
; and
(ii)
hesit
ancy
abo
ut
vacc
ines
from
spec
ific
man
ufac
ture
rs w
as n
ot o
bser
ved
or su
spec
ted
as n
o re
spon
dent
men
tione
d th
e na
me
of a
vac
cine
– th
ey re
ferre
d to
the
vacc
ine
acco
rdin
g to
the
time
at w
hich
it n
eede
d to
be
adm
inist
ered
(e.g
. “va
ccin
e fo
r 2.5
mon
ths”
)
Posit
ive:
hes
itanc
y ab
out v
acci
nes f
rom
spec
ific
man
ufac
ture
rs w
as n
ot o
bser
ved
or su
spec
ted
as n
o re
spon
dent
men
tione
d th
e na
me
of a
vac
cine
– th
ey re
ferre
d to
the
vacc
ine
acco
rdin
g to
the
time
at w
hich
it n
eede
d to
be
adm
inist
ered
(e.g
. “v
acci
ne fo
r 2.
5 m
onth
s”)
Trus
t in,
and
per
sona
l exp
erie
nce
of, t
he h
ealth
syst
em a
nd h
ealth
pr
ofes
siona
ls
Posit
ive:
(i) h
ealth
cen
tre st
aff (m
ostly
nur
ses a
nd v
acci
nato
rs) a
re se
en a
s a g
ood
sour
ce o
f inf
orm
atio
n (4
0 re
spon
dent
s) a
nd a
re c
onta
cted
to d
iscus
s que
stio
ns o
n va
ccin
atio
n (2
9 re
spon
dent
s); a
nd
(ii) r
espo
nden
ts m
entio
ned
that
get
ting
info
rmat
ion
on o
ther
hea
lth to
pics
(e.g
. stu
ntin
g) d
urin
g th
e va
ccin
atio
n se
ssio
ns a
nd re
ceiv
ing
shish
a ki
bond
o (i.
e. p
orrid
ge) w
ere
addi
tiona
l ben
efits
Posit
ive:
hea
lth c
entre
staff
(mos
tly n
urse
s and
vac
cina
tors
) mos
t fre
quen
tly
men
tione
d as
a so
urce
of i
nfor
mat
ion
(69
resp
onde
nts)
and
as a
con
tact
for
disc
ussin
g qu
estio
ns (5
5 re
spon
dent
s).
Neg
ativ
e: so
me
mot
hers
had
neg
ativ
e ex
perie
nces
whe
n th
eir c
hild
was
not
gi
ven
the
vacc
ine
(e.g
. the
y w
ere
sent
hom
e be
caus
e th
ey w
ere
late
or b
ecau
se a
m
ultid
ose
vial
cou
ld n
ot b
e op
ened
) or t
hey
rece
ived
aw
fine
(e.g
. for
bei
ng la
te,
not h
avin
g th
eir c
hild
vac
cina
ted
or n
ot re
-eng
agin
g w
ith th
e va
ccin
atio
n sy
stem
af
ter a
hom
e de
liver
y)Co
mpl
acen
cy
Co
mm
unic
atio
n an
d m
edia
en
viro
nmen
tN
eutra
l: (i)
radi
o is
still
a so
urce
of i
nfor
mat
ion
on v
acci
natio
n se
rvic
es (3
1 re
spon
dent
s); (
ii) g
over
nmen
t ca
mpa
ign
mat
eria
ls (e
.g. fl
yers
) wer
e m
entio
ned
less
ofte
n as
a so
urce
of i
nfor
mat
ion
(2 re
spon
dent
s);
and
(iii)
inte
rest
ingl
y, th
e va
ccin
atio
n ca
rd w
as a
lso e
xplic
itly
men
tione
d as
a so
urce
of i
nfor
mat
ion
(10
resp
onde
nts)
Neu
tral:
(i) ra
dio
is st
ill a
sour
ce o
f inf
orm
atio
n on
vac
cina
tion
serv
ices
(35
resp
onde
nts)
; and
(ii)
gove
rnm
ent c
ampa
ign
mat
eria
ls (e
.g. fl
yers
) wer
e fre
quen
tly m
entio
ned
as a
sour
ce o
f inf
orm
atio
n (2
0 re
spon
dent
s), a
s wer
e va
ccin
atio
n ca
rds (
6 re
spon
dent
s)In
fluen
tial l
eade
rs, i
mm
uniz
atio
n pr
ogra
mm
e ga
teke
eper
s and
va
ccin
atio
n lo
bbie
s
Posit
ive:
no
care
give
rs m
entio
ned
loca
l lea
ders
as a
neg
ativ
e in
fluen
ce (o
ne re
spon
dent
repo
rted
how
th
e lo
cal l
eade
r goe
s aro
und
the
villa
ge w
ith a
loud
spea
ker t
o gi
ve in
form
atio
n ab
out v
acci
natio
n an
d ot
her a
ctiv
ities
)
Non
e
Hist
oric
al in
fluen
ces
Posit
ive:
mea
sles o
utbr
eak
in 2
019
incr
ease
d th
e nu
mbe
r of v
isibl
e ca
ses
Non
eRe
ligio
n, c
ultu
re, g
ende
r and
so
cioe
cono
mic
fact
ors
Posit
ive:
resp
onde
nts m
entio
ned
that
, whe
n th
ey a
re si
ck, t
here
are
peo
ple
in th
e co
mm
unity
(e.g
. hu
sban
d, C
HW
or n
eigh
bour
) who
can
take
the
child
for v
acci
natio
n (i.
e. c
omm
unity
eng
agem
ent).
N
egat
ive:
(i) c
areg
iver
s men
tione
d th
at a
few
car
egiv
ers d
o no
t brin
g th
eir c
hild
ren
to th
e va
ccin
atio
n se
ssio
n, d
escr
ibin
g th
em a
s “ca
rele
ss” (
3 re
spon
dent
s) o
r “bu
sy w
ith li
fe” (
1 re
spon
dent
); an
d (ii
) fam
ily
confl
ict (
1 re
spon
dent
) and
chi
ldre
n w
ho c
ry d
urin
g th
e ni
ght a
fter v
acci
natio
n (1
resp
onde
nt) w
ere
men
tione
d as
reas
ons f
or n
ot a
ttend
ing
vacc
inat
ion
Posit
ive:
som
e re
spon
dent
s men
tione
d th
at b
eing
sick
was
a b
arrie
r (4
resp
onde
nts)
but
oth
ers (
4 re
spon
dent
s) sa
id th
at, e
ven
whe
n th
ey a
re si
ck, t
here
ar
e pe
ople
in th
e co
mm
unity
(e.g
. hus
band
, CH
W o
r nei
ghbo
ur) w
ho c
an ta
ke
the
child
for v
acci
natio
n (i.
e. c
omm
unity
eng
agem
ent).
N
egat
ive:
som
e ca
regi
vers
said
that
oth
ers m
issed
app
oint
men
ts d
ue to
“c
arel
essn
ess”
(11
resp
onde
nts)
, wor
k (2
resp
onde
nts)
or g
ivin
g bi
rth
at h
ome
(6 re
spon
dent
s) b
ut th
ey o
ften
belie
ved
they
wou
ld c
ome
late
r (e.
g. a
fter a
visi
t fro
m a
CH
W)
(contin
ues.
. .)
787Bull World Health Organ 2021;99:783–794D| doi: http://dx.doi.org/10.2471/BLT.20.285258
ResearchVaccine hesitancy mechanisms, RwandaCatherine Decouttere et al.
Fact
or a
ffect
ing
vacc
ine
hesit
ancy
a
Sele
cted
com
men
ts o
n fa
ctor
by c
areg
iver
sb,c,d
,e
Rura
l hea
lth ce
ntre
s in
Ram
ba a
nd Su
vof (n
= 74
)Pe
ri-ur
ban
heal
th ce
ntre
in G
ahan
gag (n
= 87
)
Know
ledg
e an
d aw
aren
ess
Posit
ive:
(i) C
HW
s wer
e m
entio
ned
mos
t fre
quen
tly a
s a so
urce
of i
nfor
mat
ion
on v
acci
natio
n se
rvic
es
(67
resp
onde
nts)
and
as a
con
tact
for a
skin
g qu
estio
ns a
bout
serv
ices
(70
resp
onde
nts)
but
they
wer
e al
so th
ough
t im
port
ant f
or c
omm
unity
mob
iliza
tion
and
follo
w-u
p (4
resp
onde
nts)
; (ii)
CH
W fo
llow
-up
was
exp
licitl
y m
entio
ned
as h
avin
g im
prov
ed o
ver t
he y
ears
(3 re
spon
dent
s) –
“In
the
past
, you
cou
ld
even
not
fini
sh a
ll va
ccin
es/a
ppoi
ntm
ents
and
ther
e w
as n
o on
e to
follo
w-u
p on
you
but
if y
ou d
o no
t co
me
as p
er y
our a
ppoi
ntm
ent i
n th
ese
days
, a C
HW
will
reac
h ou
t to
you
and
ask
why
you
did
not
go
for v
acci
natio
n an
d ad
vise
you
on
how
to c
atch
up”
; and
(iii)
oth
er c
omm
unity
mem
bers
(mos
tly o
ther
m
othe
rs a
nd n
eigh
bour
s) w
ere
frequ
ently
men
tione
d as
sour
ces o
f inf
orm
atio
n on
vac
cina
tion
serv
ices
(1
6 re
spon
dent
s)
Posit
ive:
(i) r
espo
nden
ts m
entio
ned
that
com
mun
ity m
embe
rs w
ere
awar
e of
th
e im
port
ance
of v
acci
natio
n be
caus
e of
com
mun
ity m
obili
zatio
n, in
volv
ing,
for
exam
ple,
CH
Ws a
nd c
ampa
igns
; and
(ii)
CHW
s wer
e m
entio
ned
as th
e se
cond
m
ost f
requ
ent s
ourc
e of
info
rmat
ion
on v
acci
natio
n se
rvic
es (4
5 re
spon
dent
s)
and
as a
con
tact
for a
skin
g qu
estio
ns a
bout
thes
e se
rvic
es (4
4 re
spon
dent
s)
Perc
eive
d ris
ks a
nd b
enefi
tsPo
sitiv
e: (i
) the
nee
d fo
r dise
ase
prev
entio
n w
as st
reng
then
ed b
y kn
owle
dge
of c
ases
of i
llnes
s or d
eath
du
e to
vac
cine
-pre
vent
able
dise
ases
(25
resp
onde
nts)
; and
(ii)
all r
espo
nden
ts w
ere
high
ly m
otiv
ated
to
atte
nd v
acci
natio
n se
ssio
ns b
y th
eir d
esire
to p
reve
nt d
iseas
e or
ens
ure
thei
r chi
ldre
n w
ill g
row
up
to b
e he
alth
y
Posit
ive:
in a
dditi
on to
thei
r rol
e in
pre
vent
ing
dise
ase,
vac
cine
s wer
e tru
sted
be
caus
e th
ey h
ad h
ad n
o ne
gativ
e eff
ects
so fa
r (7
resp
onde
nts)
. N
egat
ive:
(i) f
ear o
f adv
erse
effe
cts f
ollo
win
g im
mun
izat
ion
(2 re
spon
dent
s); a
nd
(ii) m
ost r
espo
nden
ts d
id n
ot k
now
of c
ases
of i
llnes
s or d
eath
due
to v
acci
ne-
prev
enta
ble
dise
ases
– k
now
n ca
ses (
7 re
spon
dent
s) w
ere
mos
tly p
olio
or
mea
sles,
ofte
n in
old
er p
eopl
eIm
mun
izat
ion
as a
soci
al n
orm
Posit
ive:
resp
onde
nts w
ho d
id n
ot k
now
of a
ny c
ases
of i
llnes
s or d
eath
due
to v
acci
ne-p
reve
ntab
le
dise
ases
(38
resp
onde
nts)
men
tione
d th
at th
e co
mm
unity
and
thei
r par
ents
kne
w th
e im
port
ance
of
vacc
ines
for p
rote
ctin
g ch
ildre
n; (i
i) va
ccin
atio
n se
emed
to b
e st
anda
rd (e
.g. “e
very
kid
is v
acci
nate
d in
th
e co
mm
unity
”); a
nd (i
ii) v
acci
natio
n w
as fr
eque
ntly
end
orse
d by
com
mun
ity m
embe
rs, s
uch
as fr
iend
s an
d fa
mily
Non
e
Conv
enie
nce
Avai
labi
lity
of th
e im
mun
izat
ion
serv
ice
(incl
udin
g va
ccin
e av
aila
bilit
y)
Posit
ive:
(i) o
nly
one
mot
her m
entio
ned
the
avai
labi
lity
of v
acci
nes a
s an
expl
icit
reas
on fo
r com
ing
to
the
vacc
inat
ion
sess
ion;
and
(ii)
stoc
k-ou
ts w
ere
not m
entio
ned
as a
bar
rier.
Neg
ativ
e: n
ot a
ll va
ccin
es w
ere
avai
labl
e at
out
reac
h po
sts (
one
resp
onde
nt sa
id th
at th
e va
ccin
e gi
ven
whe
n th
e ch
ild is
15
mon
ths o
f age
was
not
ava
ilabl
e)
Neg
ativ
e: c
areg
iver
s can
be
sent
aw
ay if
they
do
not h
ave
an a
ppoi
ntm
ent,
are
late
or i
f the
des
ired
vacc
ine
or a
ntig
en is
not
offe
red
that
day
, for
exa
mpl
e, to
sa
ve m
ultid
ose
vial
s
Affor
dabi
lity
of th
e im
mun
izat
ion
serv
ice
Neg
ativ
e: (i
) mor
e fin
anci
al su
ppor
t req
uest
ed (4
resp
onde
nts)
; and
(ii)
rew
ard
for b
eing
fully
vac
cina
ted
requ
este
d (1
resp
onde
nt)
Neg
ativ
e: c
ost o
f tra
nspo
rt is
a b
arrie
r (3
resp
onde
nts)
Geo
grap
hica
l acc
essib
ility
Neg
ativ
e: (i
) roa
d to
the
heal
th c
entre
is b
ad a
nd h
eavy
rain
fall
mak
es it
diffi
cult
and
risky
to a
cces
s the
va
ccin
atio
n se
rvic
e (1
2 re
spon
dent
s); (
ii) d
istan
ce w
as m
entio
ned
as a
bar
rier o
r as a
pro
blem
that
mus
t be
tack
led
(6 re
spon
dent
s), a
lthou
gh th
e ne
w h
ealth
cen
tre a
t Sov
u an
d ou
treac
h im
prov
ed th
is fo
r so
me
mot
hers
(5 re
spon
dent
s); (
iii) o
ne m
othe
r sai
d th
at o
utre
ach
post
s sho
uld
be k
ept o
pen
beca
use
they
wer
e cl
osed
for a
whi
le; a
nd (i
v) o
ne m
othe
r men
tione
d th
at v
acci
natio
n se
rvic
es w
ere
clos
e, th
anks
to
out
reac
h, b
ut th
ey st
ill n
eede
d to
wal
k fo
r mor
e th
an a
n ho
ur to
all
othe
r ser
vice
s (e.
g. to
giv
e bi
rth)
Posit
ive:
vac
cina
tion
sess
ions
wer
e ea
sy to
reac
h an
d tra
vel t
imes
wer
e sh
ort b
ut
som
e m
othe
rs st
ill re
ques
ted
addi
tiona
l out
reac
h (5
resp
onde
nts)
. N
egat
ive:
hea
vy ra
infa
ll m
akes
it d
ifficu
lt to
atte
nd (1
5 re
spon
dent
s)
Abili
ty to
und
erst
and
(i.e.
lang
uage
an
d he
alth
lite
racy
)Po
sitiv
e: o
ne m
othe
r men
tione
d sh
e ca
nnot
read
or w
rite
and
asks
fam
ily m
embe
rs o
r nei
ghbo
urs t
o re
ad th
e da
ta o
n th
e va
ccin
atio
n ca
rd.
Neg
ativ
e: fe
ar o
f bei
ng fi
ned
for a
non
-inst
itutio
nal b
irth
desp
ite th
e fin
e no
lo
nger
exi
stin
g (1
resp
onde
nt)
(. . .continued)
(contin
ues.
. .)
788 Bull World Health Organ 2021;99:783–794D| doi: http://dx.doi.org/10.2471/BLT.20.285258
ResearchVaccine hesitancy mechanisms, Rwanda Catherine Decouttere et al.
was negatively impacted by the supply of measles and rubella vaccine being limited outside of health centres due to a desire to reduce wastage of multi-dose vials. In addition, poor families, particularly in rural areas, were unable to attend sessions at health centres because of travel costs, distance, safety or the time required. In peri-urban settings and at district and central levels, the efficiency of, and the level of attendance at, outreach services were reported to benefit from the support of CHWs. As well as forgetting the second vaccine dose at 15 months, experienced mothers associated the dispensing of mosquito nets at 9 months (at the time of the first dose) with the end of childhood immunization, even after the second dose had been introduced. The main factor mentioned by Expanded Programme on Immunization and peri-urban health centre staff as having a negative effect on convenience was a long waiting time. In rural settings, time-saving strategies such as holding vaccination sessions on market days were regarded as having a positive impact on convenience and also on the provision of nutritional and family planning services. Finally, interviewees from all levels and settings said that paper-based data management made it difficult to monitor late immunizations and drop-outs (i.e. vaccine defaulters).
Caregivers’ perspective
Factors associated with vaccine hesitan-cy reported in interviews with caregivers in Ramba, Sovu and Gahanga were also categorized using the 3Cs framework (Table 3). With regard to confidence, caregivers reported no concerns about vaccine quality and trusted vaccines. Additionally, in rural settings, trust was reported to stem from respect for providers, including nurses and CHWs. CHWs were more often reported as a source of information in rural than in peri-urban settings. The impact of gov-ernment information campaigns seemed greater in peri-urban settings.
With regard to complacency, caregiv-ers did not mention misinformation on social media in either setting. Carelessness and forgetting were considered to be the main reasons for underimmunization, particularly when mothers had older chil-dren, more tasks and different priorities.
Factors affecting convenience var-ied considerably between rural and peri-urban settings, with differences in travel Fa
ctor
affe
ctin
g va
ccin
e he
sitan
cya
Sele
cted
com
men
ts o
n fa
ctor
by c
areg
iver
sb,c,d
,e
Rura
l hea
lth ce
ntre
s in
Ram
ba a
nd Su
vof (n
= 74
)Pe
ri-ur
ban
heal
th ce
ntre
in G
ahan
gag (n
= 87
)
Qua
lity
of th
e se
rvic
e (p
erce
ived
or
real
)Po
sitiv
e: th
e m
ajor
ity o
f car
egiv
ers w
ere
satis
fied
with
the
curre
nt se
rvic
e an
d di
d no
t sug
gest
cha
nges
or
add
ition
s (41
resp
onde
nts)
. N
eutra
l: su
gges
tions
for i
mpr
ovem
ents
rela
ted
to o
ther
serv
ices
, suc
h as
pro
vidi
ng v
itam
ins (
1 re
spon
dent
) and
pro
vidi
ng sh
isha
kibo
ndo
(por
ridge
) to
all c
hild
ren
and
not o
nly
to so
me,
as t
his i
s pe
rcei
ved
as u
nfai
r (4
resp
onde
nts)
. One
resp
onde
nt sa
id: “
Due
to fa
mily
pla
nnin
g, re
cent
dev
elop
men
t an
d se
nsiti
zatio
n, th
ose
who
did
not
resp
ect f
amily
pla
nnin
g an
d ke
pt g
ivin
g bi
rth
or d
eliv
erin
g ar
e al
way
s rec
eivi
ng fo
rtifi
ed fo
ods o
r foo
d su
pple
men
ts w
hile
, for
us w
ho re
spec
ted
it, a
re so
meh
ow
puni
shed
and
do
not g
et a
cces
s to
food
supp
lem
ents
and
yet
we
still
hav
e ch
ildre
n. O
ur w
ish a
nd
requ
est i
s to
chan
ge th
is de
cisio
n an
d gi
ve fo
rtifi
ed fo
ods t
o al
l of u
s ins
tead
of p
unish
ing
us”
Non
e
Conv
enie
nt ti
me
(incl
udin
g w
aitin
g tim
e), p
lace
and
cul
tura
l con
text
Posit
ive:
a d
esire
to re
spec
t the
giv
en a
ppoi
ntm
ent i
s the
seco
nd m
ost f
requ
ent r
easo
n fo
r atte
ndin
g a
spec
ific
sess
ion
(41
resp
onde
nts)
afte
r the
des
ire to
“pro
tect
chi
ldre
n ag
ains
t dise
ases
,” whi
ch su
gges
ts
that
the
appo
intm
ent s
yste
m, a
ided
by
vacc
inat
ion
card
s, is
effec
tive.
N
egat
ive:
(i) c
lean
wat
er, e
lect
ricity
or b
oth
requ
este
d at
the
heal
th c
entre
(2 re
spon
dent
s); (
ii) lo
ng
wai
ting
times
(2 re
spon
dent
s); a
nd (i
ii) th
e ne
ed to
redu
ce w
aitin
g tim
es c
ompa
red
with
pre
viou
s se
ssio
ns (2
resp
onde
nts)
Neg
ativ
e: (i
) lon
g w
aitin
g tim
e du
e to
too
few
nur
ses w
as a
big
issu
e (4
2 re
spon
dent
s); (
ii) a
s a re
sult,
the
wai
ting
room
was
cro
wde
d an
d pe
ople
had
to
wai
t out
side;
and
(iii)
one
resp
onde
nt su
gges
ted
that
the
mor
ning
hou
rs a
re
bette
r as i
t is l
ess h
ot to
wal
k w
ith a
chi
ld to
the
heal
th c
entre
Des
ign
of v
acci
natio
n pr
ogra
mm
e an
d va
ccin
atio
n sc
hedu
leN
eutra
l: m
othe
rs w
ith m
ore
than
one
chi
ld m
entio
ned
that
the
vacc
inat
ion
sche
dule
and
vac
cine
s had
ch
ange
d sin
ce th
ey h
ad th
eir fi
rst c
hild
(e.g
. an
extra
vac
cine
is n
eede
d w
hen
the
child
is 1
5 m
onth
s old
an
d tw
o or
thre
e in
ject
ions
are
requ
ired
inst
ead
of o
ne o
r tw
o)
Non
e
CHW
: com
mun
ity h
ealth
wor
ker.
a Fac
tors
wer
e ca
tego
rized
usin
g th
e 3C
s fra
mew
ork
as re
late
d to
con
fiden
ce, c
ompl
acen
cy o
r con
veni
ence
.8
b Com
men
ts w
ere
clas
sified
as p
ositi
ve, n
eutra
l or n
egat
ive
with
rega
rd to
thei
r im
plic
atio
ns fo
r vac
cine
upt
ake.
c The
num
bers
in th
e ta
ble
refe
r to
the
num
ber o
f int
ervi
ewee
s who
agr
eed
with
the
com
men
t.d C
hild
ren’s
car
egiv
ers w
ere
inte
rvie
wed
dur
ing
vacc
inat
ion
sess
ions
at h
ealth
cen
tres a
nd o
utre
ach
post
s in
two
com
mun
ities
.e D
etai
ls of
whi
ch in
terv
iew
ees m
ade
each
com
men
t are
ava
ilabl
e fro
m th
e da
ta re
posit
ory.32
f The
Ram
ba h
ealth
cen
tre is
in N
goro
rero
Dist
rict.
Afte
r a m
easle
s out
brea
k in
201
9, o
ne o
f the
hea
lth c
entre
’s ou
treac
h po
sts b
ecam
e th
e So
vu h
ealth
cen
tre.
g The
Gah
anga
hea
lth c
entre
is in
the
Kicu
kiro
Dist
rict o
f the
Mun
icip
ality
of K
igal
i.
(. . .continued)
789Bull World Health Organ 2021;99:783–794D| doi: http://dx.doi.org/10.2471/BLT.20.285258
ResearchVaccine hesitancy mechanisms, RwandaCatherine Decouttere et al.
distances and waiting times. In the rainy season, travelling safely with young chil-dren was complicated in rural areas and immunization was delayed; some people relied completely on outreach services.
Comparing perspectives
Both service providers and caregivers largely agreed on the strengths of the vac-cination programme but identified differ-ent challenges (Box 1, based on Table 2 and Table 3). By considering caregivers’ insights, immunization providers were able to obtain a broader perspective on their services, which in turn provided a basis for designing future interventions. For example, the introduction of an ef-ficient digital data management system could help tackle the multiple challenges perceived by immunization providers while maintaining an appointment sys-tem well regarded by caregivers.
Causal loop diagrams
The causal relationships between the main factors affecting vaccine hesitancy identified in interviews with vaccination service providers and caregivers are il-lustrated in three causal loop diagrams in Fig. 2 (available at https:// www .who .int/ publications/ journals/ bulletin/), for confidence, complacency and con-venience, respectively. Fig. 3, which is a composite of these three diagrams, indicates that vaccine uptake is governed by three key factors: (i) trust in vaccina-tion; (ii) community engagement; and (iii) access to vaccination. As vaccine uptake evolves, six feedback loops are activated (Fig. 3): three balancing B loops and three reinforcing R loops that can increase or decrease trust, engage-ment or access. These loops illustrate the dynamic nature of the system. The same factors and loops were identified in both rural and peri-urban settings but their relative impact on vaccination uptake differed. For example, in rural settings, access to vaccinations was reduced by a lack of outreach services and information campaigns, whereas, in peri-urban settings, workload pres-sure on the immunization system had a negative impact on trust in vaccination.
To maintain measles vaccination coverage at the desired level of over 95%, behaviour that increases vaccine uptake, as indicated by loops in the causal loop diagram (Fig. 3), must be actively pro-moted by health interventions. For ex-
ample, government-led information and vaccination campaigns, community ad-vocacy and direct communication from CHWs during home visits all regularly boost the perceived benefit of vaccina-tion and community engagement. The disruption caused by the coronavirus disease 2019 (COVID-19) pandemic affected the 3Cs: (i) convenience was diminished because vaccination ses-sions were smaller and less accessible; (ii) confidence was reduced by height-ened pressure on the system and fear of infection at the health centre; and (iii) complacency was increased by less contact between CHWs and caregivers.
The composite causal loop dia-gram can also be used to illustrate the circumstances that led to the 2019 measles outbreak in Sovu (Fig. 4). First, community and caregiver engagement with the second measles and rubella immunization dose for children aged
15 months was low because mothers had other priorities, perceived their children as stronger due to being older, or were not aware of the second dose. Second, in the absence of outreach services, health centres did not know which children were underimmunized. Therefore, the need for an additional immunization campaign or outreach services was not recognized. The resulting outbreak trig-gered a new immunization campaign and further investigations. Eventually, a new health centre was established in the affected area, which reduced the previously underimmunized popula-tion’s dependency on both outreach services and immunization campaigns. The ongoing development of digital data management systems at health centres will support this change. In addition, CHWs can help bridge the gap between caregivers and the health system and can assist in surveillance.
Box 1. Vaccination service providers’ and caregivers’ perspectives on the strengths of, and challenges to, the measles vaccination programme in rural and peri-urban Rwanda, 2018–2020
StrengthsAll interviewees:
• CHWs provide information, mobilize caregivers and follow up vaccines;
• health centre staff provide information;
• confidence in vaccination is high;
• there is an intrinsic motivation to improve health; and
• community support.
Vaccination service providers:
• health ministry committed to supporting CHWs and mother-and-child health weeks.
Caregivers:
• vaccination cards;
• information provided by radio;
• high service quality; and
• appointments system.
ChallengesAll interviewees:
• outreach services needed for poor and hard-to-reach families.
Vaccination service providers:
• high staff turnover and workload and need for training;
• tracing of vaccine defaulters inefficient;
• outreach services for measles and rubella vaccination inadequate; and
• mothers missing the second measles and rubella vaccine dose for their children.
Caregivers:
• difficult vaccination access in the rainy season;
• long distances to health centres; and
• long waiting times.
CHW: community health worker.
790 Bull World Health Organ 2021;99:783–794D| doi: http://dx.doi.org/10.2471/BLT.20.285258
ResearchVaccine hesitancy mechanisms, Rwanda Catherine Decouttere et al.
Fig. 3. Composite causal loop diagram of factors affecting vaccine hesitancy, Rwanda, 2018–2020
Pressure on immunization
system
Disease incidence
–
R: Reinforcing loop B: Balancing loop
+
+
+
+
+
+
Community engagement
Perceived vaccine benefit
Vaccination–
+ Positive effect – Negative effect
–+
++
Relationship with health centre staff
Perceived quality of vaccination service
Trust in vaccination
Perceived threat of disease
R Complacency
B Complacency
R Confidence
BConfidence
Outreach and campaign planning
Outreach and campaign need
Individual health
R Convenience
B Convenience
Access to vaccination
Health literacy and mobility
Protected population+
+
+
+
–
+
+
+
Notes: This diagram is the composite of the three causal loop diagrams in Fig. 2. Factors affecting vaccine hesitancy were categorized using the 3Cs framework as related to confidence, complacency or convenience.8 The R loops are reinforcing feedback loops that cause change in the same direction and the B loops are balancing loops that cause change in the opposite direction.
Fig. 4. Composite causal loop diagram of factors affecting vaccine hesitancy during a measles outbreak, Rwanda, 2019
Pressure on immunization
system
Disease incidence
–
R: Reinforcing loop B: Balancing loop
+
+
+
+
+
+
Community engagement
Perceived vaccine benefit
Vaccination–
+ Positive effect – Negative effect
-+
+
+
Relationship with health centre staff
Perceived quality of vaccination service
Trust in vaccination
Perceived threat of disease
R Complacency
B Complacency
R Confidence
BConfidence
Outreach and campaign planning
Outreach and campaign need
Individual health
R Convenience
B Convenience
Access to vaccination
Health literacy and mobility
Protected population+
+
+
+
–
+
+
+
Surveillance
Child ageing
+
+
Caregiver’sother priorities
–
Communication campaign
Community health workers
Digital data management
New health centre
+
+
+
+
–+
Important factors influencing vaccination uptake Interventions introduced to decrease vaccine hesitancy
Notes: The measles outbreak occurred in the catchment area of the Ramba health centre in 2019. This diagram is an adaptation of the causal loop diagram in Fig. 3. Factors affecting vaccine hesitancy were categorized using the 3Cs framework as related to confidence, complacency or convenience.8 The R loops are reinforcing feedback loops that cause change in the same direction and the B loops are balancing loops that cause change in the opposite direction.
791Bull World Health Organ 2021;99:783–794D| doi: http://dx.doi.org/10.2471/BLT.20.285258
ResearchVaccine hesitancy mechanisms, RwandaCatherine Decouttere et al.
DiscussionOur analysis of factors influencing vac-cine hesitancy in Rwanda offers several insights. For example, trust in vaccina-tion and social cohesion are factors that can be leveraged in various settings. However, the differences we identified between rural and peri-urban settings in the ease of travel indicate that solutions for vaccine hesitancy and vaccine policy design are dependent on the setting, even for the same vaccine in the same country. Both vaccination service providers (i.e. Expanded Programme on Immunization staff and CHWs) and caregivers thought that the accessibility of vaccination ses-sions and the quality of the immunization service were important influences behind underimmunization. However, the in-sights we obtained from service providers and caregivers revealed they had differing perspectives, which provided an op-portunity for collective learning and for increasing vaccine uptake.
Some service delivery policies had unintended consequences. For instance, according to national guidelines, multi-dose vaccine vials must be used at fixed sites rather than at both outreach posts and fixed sites to reduce wastage. This policy may increase vaccine hesitancy by reducing access to, or the convenience of, vaccination for people dependent on outreach services. Moreover, in some places in Rwanda, there used to be a financial penalty if a child was not reg-istered with the vaccine system at birth. Consequently, families whose children were born at home were often deterred from visiting health centres due to mis-information that the penalty still existed. Conversely, this lasting fear of a financial penalty may incentivize some pregnant mothers to engage with vaccination.
Our findings illustrate the crucial role of community engagement in building system resilience. Causal loops can be strengthened or weakened by external elements, such as interven-tions made in response to a measles outbreak. For instance, communication within communities can be leveraged by prolonging CHW programmes, thereby enhancing social cohesion. In contrast, the COVID-19 pandemic induced the perception that health centre visits were unsafe. This perception, combined with
the temporary cancellation of commu-nity awareness activities and educational sessions at health centres, resulted in immunization being delayed until mid-August 2020. The speed at which vaccine uptake was restored after the disruption showed the resilience of the system.35,36
As outreach services are a cor-nerstone of measles immunization programmes in some populations, their sustainability will affect future immu-nization programmes. The efficiency of outreach could be increased by focusing efforts on hard-to-reach populations or by establishing temporary outreach posts during the rainy season. Depen-dency on outreach services could be reduced by opening new health-care delivery points or by providing physical or financial support to enable caregivers to travel to health centres, both of which would enhance the sustainability of the immunization system.
Our study of two settings in Rwanda revealed leverage points that cut across several factors influencing vaccine hesitancy (i.e. one specific intervention can impact multiple loops within the immunization system). For example, we found that the connecting role of CHWs was pivotal and that they could function as a high-potential leverage point because they have a direct impact on two feedback loops in Fig. 3: the balancing convenience loop (i.e. iden-tifying the need for outreach and vac-cination campaigns) and the reinforc-ing complacency loop (i.e. increasing awareness of the benefits of vaccination through home visits).25 Furthermore, the digitization of local immunization data, such as the immunization status of the population (irrespective of the point of vaccination), could guide targeted and timely preventive interventions (e.g. catch-up vaccinations in the child’s second year of life), assist in campaign planning and help trace defaulters. Nev-ertheless, the factors affecting vaccine hesitancy can change over time (e.g. in response to vaccine-hesitant social media influencers).37
Our study has limitations. First, the presence of a well-functioning vaccina-tion delivery system in Rwanda may make it more difficult to generalize our findings. Second, although we carefully selected the study settings, they were
limited in number and our findings relate to only rural and peri-urban communities. More research is needed on urban and other settings. Third, we were not able to interview caregiv-ers who are not present at vaccination sessions. However, interviewees were asked about the motivations of caregiv-ers who missed sessions. Similarly, we interviewed only CHWs and other vac-cination service providers who spoke on behalf of their teams, which may have introduced bias. Finally, the COVID-19 pandemic complicated fieldwork.
In addition to overcoming these limitations, future research could build on our insights and causal loop dia-grams to develop a human-centred, collaborative approach to identifying leverage points that could be used to design sustainable interventions for minimizing vaccine hesitancy. More-over, our systems thinking approach and factors influencing hesitancy could be integrated with initiatives like the Vaccine Confidence Project,38 which contains a tool for mapping confidence globally.12,14,39,40 ■
AcknowledgementsWe thank Christine Lee, Alan Zhang, the Expanded Programme on Immuniza-tion team in Rwanda and all participants in workshops, interviews and field visits.
Funding: This work was funded by GSK through the GSK research chair on the redesign of health care supply chains in developing countries to increase access-to-medicines, KU Leuven.
Competing interests: Stany Banzimana received a PhD scholarship funded by GSK. Catherine Decouttere was funded through a GSK research chair at KU Leuven. Corinne Vandermeulen was a principal investigator on a study for GSK unrelated to this research for which KU Leuven received a grant.
792 Bull World Health Organ 2021;99:783–794D| doi: http://dx.doi.org/10.2471/BLT.20.285258
ResearchVaccine hesitancy mechanisms, Rwanda Catherine Decouttere et al.
摘要从系统思维深入了解“疫苗犹豫”,卢旺达目的 调查导致卢旺达免疫不足和麻疹爆发的“疫苗犹豫”,并开发一个社区层面的概念性行为因素模型。方法 采用系统思维、以人为本的设计理念和行为框架,探索了卢旺达两个社区(其中一个地区最近经历了麻疹爆发)的当地免疫系统。收集的 2018 年至 2020 年数据来自 :与 11 位疫苗接种服务提供者(即医院和卫生中心工作人员)的讨论 ;采访了卫生中心的 161 名儿童看护人员 ;以及与卫生中心工作人员进行的九次验证访谈。使用 3C 框架对影响“疫苗犹豫”的因素进行分类 :信心感、满足度和便利性。开发了具有反馈回路的“疫苗犹豫”机制的概念模型。结果 对农村和城郊环境中服务提供者和看护人员的观点进行比较,结果表明以下类似因素加强了疫苗接种:
(I) 基于与卫生中心工作人员的个人关系对疫苗和服务提供者的高度信任 ;(ii) 社区卫生工作者所起到的纽带作用;(iii) 强烈的社区意识。被确定为增加“疫苗犹豫”的因素(例如服务的可及性和后续随访不足)在不同的环境中以及在不同的服务提供者和看护人员之间有所差异。该概念模型可用于解释近期麻疹爆发的驱动因素,并指导提高疫苗接种率的干预措施。结论 行为框架和系统思维的应用揭示了卢旺达社区的
“疫苗犹豫”机制,证明了免疫服务与护理人员的疫苗接种行为之间的相互关系。还确定了营造信心的社会结构和不同环境下影响疫苗接种的挑战。
Résumé
Analyse de l'hésitation vaccinale à l'aide de la pensée systémique au Rwanda Objectif Analyser l'hésitation vaccinale menant à une immunisation insuffisante et à une épidémie de rougeole au Rwanda, mais aussi développer un modèle conceptuel de facteurs comportementaux à l'échelle communautaire.Méthodes Nous avons examiné les systèmes d'immunisation locaux dans deux communautés rwandaises (l'une ayant récemment subi une épidémie de rougeole) en utilisant la pensée systémique, une approche centrée sur l'humain et des cadres comportementaux. Les données ont été récoltées entre 2018 et 2020 auprès de plusieurs sources: les discussions avec 11 prestataires de services liés à la vaccination (hôpitaux et centres médicaux); des entretiens avec 161 aides-soignants du département pédiatrique dans divers centres médicaux; et enfin, neuf entretiens de validation avec le personnel des centres médicaux. Nous avons classé les facteurs conditionnant l'hésitation vaccinale conformément à la règle des 3 C: confiance (confidence), sous-estimation
du danger (complacency) et commodité (convenience). Un modèle conceptuel des mécanismes d'hésitation vaccinale assorti de boucles de rétroaction a également été élaboré.Résultats D'après une comparaison entre les perspectives des prestataires de services et aides-soignants, tant en milieu rural que périurbain, des facteurs similaires renforcent la prise vaccinale: (i) un degré de confiance élevé envers les vaccins et les prestataires de services, fondé sur les relations entretenues avec le personnel des centres médicaux; (ii) le rôle d'intermédiaire joué par les agents de santé communautaires; et enfin, (iii) un fort sentiment d'appartenance à la collectivité. Les facteurs qui ont tendance à accroître l'hésitation vaccinale (parmi lesquels figurent l'accessibilité aux services et l'absence de suivi adéquat) varient entre prestataires de services et aides-soignants, ainsi que selon les milieux. Le modèle conceptuel pourrait servir à
ملخصنظرة متعمقة على التردد في تعاطي اللقاح من واقع التفكير النظامي، رواندا
يؤدي ما وهو اللقاح تعاطي في التردد عن الاستقصاء الغرض إلى نقص المناعة وتفشي مرض الحصبة في رواندا، وتطوير نموذج
مفاهيمي على مستوى المجتمع للعوامل السلوكية.الطريقة تم استكشاف أنظمة التحصين المحلية في مجتمعين روانديين (تعرض أحدهما مؤخرًا لتفشي مرض الحصبة)، باستخدام التفكير النظامي، والتصميم المرتكز على الإنسان، وأطر العمل السلوكية. مناقشات من: و2020 2018 عامي بين البيانات تجميع تم بالمستشفى العمل فريق (أي التطعيم خدمة مقدمي من 11 مع مقدمي من 161 مع شخصية ومقابلات الصحي)؛ والمركز شخصية مقابلات وتسع الصحية؛ المراكز في للأطفال الرعاية التي العوامل تم تصنيف المركز الصحي. للتحقق مع فريق عمل :3Cs عمل إطار باستخدام اللقاح تعاطي في التردد على تؤثر (Complacency)، والراحة (Confidence)، والرضا الثقة (Convenience). تم تطوير نموذج مفاهيمي لآليات التردد في
تعاطي اللقاح مع حلقات طرح التعليقات.الخدمات، ووجهات مقدمي نظر بين وجهات المقارنة إن النتائج الحضرية، وشبه الريفية المناطق من كل في الرعاية، مقدمي نظر
تعاطي نسبة من زادت قد مماثلة عوامل هناك أن أظهرت أساس على الخدمة ومقدمي اللقاحات في عالية ثقة (1) اللقاح: الدور الصحي؛ و(2) المركز فريق عمل مع الشخصية العلاقات بالانتماء (3) شعور قوي و بالمجتمع؛ الصحيين للعاملين المتصل للمجتمع. إن العوامل التي تم التوصل إلى أنها تزيد من التردد في غير والمتابعة الخدمة، على الحصول إمكانية (مثل اللقاح تعاطي الكافية)، اختلفت بين مقدمي الخدمة ومقدمي الرعاية، وفيما بين تفشي دوافع لشرح المفاهيمي النموذج استخدام يمكن المواقع. نسبة لزيادة المقررة التدخلات ولتوجيه مؤخرًا، الحصبة مرض
تعاطي اللقاح.الاستنتاج إن كل من تطبيق أطر العمل السلوكية والتفكير النظامي كشفا عن آليات التردد في تعاطي اللقاح في المجتمعات الرواندية، وسلوك التحصين خدمات بين المتبادلة العلاقة توضح والتي التطعيم لدى مقدمي الرعاية. كما تم التوصل إلى الهياكل الاجتماعية على تؤثر والتي السياق، على تعتمد التي والتحديات الثقة، لبناء
نسبة تعاطي اللقاح.
793Bull World Health Organ 2021;99:783–794D| doi: http://dx.doi.org/10.2471/BLT.20.285258
ResearchVaccine hesitancy mechanisms, RwandaCatherine Decouttere et al.
identifier les moteurs de la récente épidémie de rougeole, et orienter les actions destinées à améliorer la prise vaccinale.Conclusion L'application des cadres comportementaux et de la pensée systémique ont révélé les mécanismes qui sous-tendent l'hésitation vaccinale au sein des communautés rwandaises, et qui prouvent la
corrélation entre les services d'immunisation et le comportement des aides-soignants en la matière. Nous avons également découvert des structures sociales de restauration de la confiance ainsi que des défis qui, en fonction du contexte, ont un impact sur la prise vaccinale.
Резюме
Понимание недоверия к вакцинации на основе системного мышления, РуандаЦель Изучить недоверие к вакцинации, ведущее к недостаточной иммунизации и вспышке кори в Руанде, а также разработать концептуальную модель поведенческих факторов на общинном уровне.Методы Были изучены местные системы по иммунизации в двух руандийских общинах (одна из которых недавно пережила вспышку кори) с использованием системного мышления, дизайна, ориентированного на потребности человека, и поведенческих структур. Были собраны данные в период с 2018 по 2020 год в результате обсуждений с 11 поставщиками услуг в области вакцинации (то есть персоналом больниц и центров здравоохранения), интервью с 161 лицом, осуществляющим уход за детьми в центрах здравоохранения, и девяти оценочных интервью с персоналом центра здравоохранения. Факторы, влияющие на недоверие к вакцинации, были классифицированы с использованием «сис темы 3 фак торов» : доверие, удовлетворенность жизненной ситуацией и удобство. Была разработана концептуальная модель механизмов недоверия к вакцинации с обратной связью.Результаты Сравнение мнений поставщиков услуг и лиц, осуществляющих уход, в сельских и пригородных районах
показало, что увеличению охвата вакцинацией способствовали аналогичные факторы: (i) высокое доверие к вакцинам и поставщикам услуг, основанное на личных отношениях с персоналом центра здравоохранения; (ii) связующая роль местных медицинских работников; (iii) сильное чувство общности. Факторы, идентифицированные как растущее недоверие к вакцинации (например, доступность обслуживания и недостаточное последующее наблюдение), различались между поставщиками услуг и лицами, осуществляющими уход, а также между районами. Концептуальную модель можно использовать для объяснения причин недавней вспышки кори и для разработки мероприятий, направленных на увеличение охвата вакцинацией.Вывод Применение поведенческих структур и системного мышления выявило в руандийских общинах механизмы недоверия к вакцинации, которые демонстрируют взаимосвязь между иммунизационным обслуживанием и поведением лиц, осуществляющих уход, в отношении вакцинации. Были также определены социальные структуры, способствующие укреплению доверия, и контекстно зависимые проблемы, влияющие на охват вакцинацией.
Resumen
Reflexiones sobre la reticencia a las vacunas desde el pensamiento sistémico en RuandaObjetivo Estudiar la reticencia a las vacunas que ha causado la falta de vacunación y la aparición de un brote de sarampión en Ruanda, y desarrollar un modelo conceptual a nivel comunitario de los factores de comportamiento.Métodos Se analizaron los programas locales de vacunación en dos comunidades de Ruanda (una de las que experimentó un brote de sarampión hace poco) mediante el uso de marcos de pensamiento sistémico, de diseño centrado en el ser humano y de comportamiento. Los datos se recopilaron entre 2018 y 2020 a partir de conversaciones con 11 proveedores de servicios de vacunación (es decir, personal de hospitales y de centros sanitarios); entrevistas con 161 cuidadores de niños en centros sanitarios; y 9 entrevistas de validación con personal de centros sanitarios. Los factores que influyen en la reticencia a la vacunación se clasificaron mediante el modelo de las 3 C: confianza, complacencia y conveniencia. Asimismo, se elaboró un modelo conceptual de los mecanismos de reticencia a la vacunación que incluye ciclos de retroalimentación.Resultados Una comparación entre las perspectivas de los proveedores de servicios y de los cuidadores, tanto en entornos rurales como
periurbanos, indicó que factores similares reforzaban la adopción de las vacunas: i) la gran confianza en las vacunas y en los proveedores de servicios gracias a las relaciones personales con el personal del centro sanitario; ii) el rol de conexión de los profesionales sanitarios de la comunidad; y iii) el gran sentido de pertenencia a la comunidad. Los factores identificados que aumentan la reticencia a las vacunas (por ejemplo, la accesibilidad a los servicios y el seguimiento insuficiente) difieren entre los proveedores de servicios y los cuidadores y entre los distintos entornos. El modelo conceptual se podría utilizar para explicar los factores que influyeron en el reciente brote de sarampión y para orientar las intervenciones diseñadas con el objetivo de aumentar la adopción de vacunas.Conclusión La aplicación de marcos de comportamiento y del pensamiento sistémico reveló mecanismos de reticencia a las vacunas en las comunidades de Ruanda que evidencian la interrelación entre los servicios de vacunación y el comportamiento de vacunación de los cuidadores. También se identificaron las estructuras sociales de fomento de la confianza y los desafíos específicos del contexto que afectan a la adopción de las vacunas.
References1. Robson J, Bao J, Wang A, McAlister H, Uwizihiwe J-P, Sayinzoga F, et al.
Making sense of Rwanda’s remarkable vaccine coverage success. Int J Healthc. 2020;6(1):56. doi: http:// dx .doi .org/ 10 .5430/ ijh .v6n1p56
2. Condo J, Mugeni C, Naughton B, Hall K, Tuazon MA, Omwega A, et al. Rwanda’s evolving community health worker system: a qualitative assessment of client and provider perspectives. Hum Resour Health. 2014 Dec 13;12(1):71. doi: http:// dx .doi .org/ 10 .1186/ 1478 -4491 -12 -71 PMID: 25495237
794 Bull World Health Organ 2021;99:783–794D| doi: http://dx.doi.org/10.2471/BLT.20.285258
ResearchVaccine hesitancy mechanisms, Rwanda Catherine Decouttere et al.
3. Seruyange E, Gahutu JB, Mambo Muvunyi C, Uwimana ZG, Gatera M, Twagirumugabe T, et al. Measles seroprevalence, outbreaks, and vaccine coverage in Rwanda. Infect Dis (Lond). 2016 Nov-Dec;48(11-12):800–7. doi: http:// dx .doi .org/ 10 .1080/ 23744235 .2016 .1201720 PMID: 27386895
4. Gatera M, Bhatt S, Ngabo F, Utamuliza M, Sibomana H, Karema C, et al. Successive introduction of four new vaccines in Rwanda: high coverage and rapid scale up of Rwanda’s expanded immunization program from 2009 to 2013. Vaccine. 2016 Jun 17;34(29):3420–6. doi: http:// dx .doi .org/ 10 .1016/ j .vaccine .2015 .11 .076 PMID: 26704259
5. Salmon DA, Dudley MZ, Glanz JM, Omer SB. Vaccine hesitancy: causes, consequences, and a call to action. Vaccine. 2015 Nov 27;33 Suppl 4:D66–71. doi: http:// dx .doi .org/ 10 .1016/ j .vaccine .2015 .09 .035 PMID: 26615171
6. MacDonald NE; SAGE Working Group on Vaccine Hesitancy. Vaccine hesitancy: definition, scope and determinants. Vaccine. 2015 Aug 14;33(34):4161–4. doi: http:// dx .doi .org/ 10 .1016/ j .vaccine .2015 .04 .036 PMID: 25896383
7. Ten threats to global health in 2019. Geneva: World Health Organization; 2020. Available from: https:// www .who .int/ news -room/ spotlight/ ten -threats -to -global -health -in -2019 [cited 2020 Dec 9].
8. Report of the Sage Working Group on vaccine hesitancy. Geneva: World Health Organization; 2014. Available from: https:// www .who .int/ immunization/ sage/ meetings/ 2014/ october/ 1 _Report _WORKING _GROUP _vaccine _hesitancy _final .pdf [cited 2021 Sep 7].
9. Petit V. The behavioural drivers model: a conceptual framework for social and behaviour change programming. Amman: United Nations Children’s Fund; 2019. Available from: https:// www .unicef .org/ mena/ media/ 5586/ file/ The _Behavioural _Drivers _Model _0 .pdf %20 .pdf [cited 2021 Sep 7].
10. Brewer NT, Chapman GB, Rothman AJ, Leask J, Kempe A. Increasing vaccination: putting psychological science into action. Psychol Sci Public Interest. 2017 Dec;18(3):149–207. doi: http:// dx .doi .org/ 10 .1177/ 1529100618760521 PMID: 29611455
11. Pruyt E, Auping WL, Kwakkel JH. Ebola in West Africa: model-based exploration of social psychological effects and interventions. Syst Res Behav Sci. 2015;32(1):2–14. doi: http:// dx .doi .org/ 10 .1002/ sres .2329
12. Larson HJ, Jarrett C, Eckersberger E, Smith DMD, Paterson P. Understanding vaccine hesitancy around vaccines and vaccination from a global perspective: a systematic review of published literature, 2007–2012. Vaccine. 2014 Apr 17;32(19):2150–9. doi: http:// dx .doi .org/ 10 .1016/ j .vaccine .2014 .01 .081 PMID: 24598724
13. Dubé E, Gagnon D, Nickels E, Jeram S, Schuster M. Mapping vaccine hesitancy – country-specific characteristics of a global phenomenon. Vaccine. 2014 Nov 20;32(49):6649–54. doi: http:// dx .doi .org/ 10 .1016/ j .vaccine .2014 .09 .039 PMID: 25280436
14. Larson HJ, Jarrett C, Schulz WS, Chaudhuri M, Zhou Y, Dube E, et al.; SAGE Working Group on Vaccine Hesitancy. Measuring vaccine hesitancy: the development of a survey tool. Vaccine. 2015 Aug 14;33(34):4165–75. doi: http:// dx .doi .org/ 10 .1016/ j .vaccine .2015 .04 .037 PMID: 25896384
15. Homer JB, Hirsch GB. System dynamics modeling for public health: background and opportunities. Am J Public Health. 2006 Mar;96(3):452–8. doi: http:// dx .doi .org/ 10 .2105/ AJPH .2005 .062059 PMID: 16449591
16. Sterman JD. Learning from evidence in a complex world. Am J Public Health. 2006 Mar;96(3):505–14. doi: http:// dx .doi .org/ 10 .2105/ AJPH .2005 .066043 PMID: 16449579
17. Resolution A/RES/70/1. Transforming our world: the 2030 agenda for sustainable development. In: Seventieth United Nations General Assembly, New York, 25 September 2015. New York: United Nations; 2015. Available from: http:// www .un .org/ ga/ search/ view _doc .asp ?symbol = A/ RES/ 70/ 1 & Lang = E [cited 2021 Sep 7].
18. Immunization Agenda 2030: a global strategy to leave no one behind. Geneva: World Health Organization; 2020. Available from: https:// www .who .int/ publications/ m/ item/ immunization -agenda -2030 -a -global -strategy -to -leave -no -one -behind [cited 2021 Sep 7].
19. Wellcome Global Monitor – first wave findings. London: Wellcome Trust; 2019. Available from: https:// wellcome .org/ sites/ default/ files/ wellcome -global -monitor -2018 .pdf [cited 2021 Sep 7].
20. Mina MJ, Kula T, Leng Y, Li M, de Vries RD, Knip M, et al. Measles virus infection diminishes preexisting antibodies that offer protection from other pathogens. Science. 2019 Nov 1;366(6465):599–606. doi: http:// dx .doi .org/ 10 .1126/ science .aay6485 PMID: 31672891
21. Petrova VN, Sawatsky B, Han AX, Laksono BM, Walz L, Parker E, et al. Incomplete genetic reconstitution of B cell pools contributes to prolonged immunosuppression after measles. Sci Immunol. 2019 Nov 1;4(41):eaay6125. doi: http:// dx .doi .org/ 10 .1126/ sciimmunol .aay6125 PMID: 31672862
22. Global measles and rubella strategic plan: 2012. Geneva: World Health Organization; 2012. Available from: https:// www .who .int/ publications/ i/ item/ 9789241503396 [cited 2021 Sep 7].
23. WHO vaccine-preventable diseases: monitoring system. 2020 global summary. WHO UNICEF estimates time series for Rwanda. Geneva: World Health Organization; 2020. Available from: https:// apps .who .int/ immunization _monitoring/ globalsummary/ estimates ?c = RWA [cited 2020 Dec 9].
24. WHO vaccine-preventable diseases: monitoring system. 2020 global summary. Incidence time series for Rwanda. Geneva: World Health Organization; 2020. Available from: https:// apps .who .int/ immunization _monitoring/ globalsummary/ incidences ?c = RWA [cited 2020 Dec 28].
25. Decouttere CJA, Vandaele N, De Boeck K, Banzimana S. A systems-based framework for immunisation system design: six loops, three flows, two paradigms [preprint]. medRxiv. 2021 July 30:2021.07.19.21260775. doi: http:// dx .doi .org/ 10 .1101/ 2021 .07 .19 .21260775
26. de Savigny D, Adam T, editors. Systems thinking for health systems strengthening. Geneva: World Health Organization; 2009. Available from: https:// www .who .int/ alliance -hpsr/ resources/ 9789241563895/ [cited 2021 Sep 7].
27. Semwanga AR, Nakubulwa S, Adam T. Applying a system dynamics modelling approach to explore policy options for improving neonatal health in Uganda. Health Res Policy Syst. 2016 May 4;14(1):35. doi: http:// dx .doi .org/ 10 .1186/ s12961 -016 -0101 -8 PMID: 27146327
28. Luke DA, Stamatakis KA. Systems science methods in public health: dynamics, networks, and agents. Annu Rev Public Health. 2012 Apr;33(1):357–76. doi: http:// dx .doi .org/ 10 .1146/ annurev -publhealth -031210 -101222 PMID: 22224885
29. Hovmand PS. Community based system dynamics. New York: Springer; 2014. doi: http:// dx .doi .org/ 10 .1007/ 978 -1 -4614 -8763 -0
30. Ozawa S, Paina L, Qiu M. Exploring pathways for building trust in vaccination and strengthening health system resilience. BMC Health Serv Res. 2016 Nov 15;16(S7) Suppl 7:639. doi: http:// dx .doi .org/ 10 .1186/ s12913 -016 -1867 -7 PMID: 28185595
31. Behavioral and social drivers of vaccination (BeSD) [internet]. Vaccination Demand Hub; 2021. Available from: https:// www .demandhub .org/ besd/ [cited 2021 Sep 7].
32. Decouttere C. Revealing vaccine hesitancy mechanisms using systems thinking: supplementary data. London: Figshare; 2021. doi: http:// dx .doi .org/ 10 .6084/ m9 .figshare .16578242 .v1
33. Poverty mapping report 2013–2014. Kigali: National Institute of Statistics of Rwanda; 2017. Available from: https:// www .statistics .gov .rw/ publication/ poverty -mapping -report -2013 -2014 [cited 2021 Sep 7].
34. Tong A, Sainsbury P, Craig J. Consolidated criteria for reporting qualitative research (COREQ): a 32-item checklist for interviews and focus groups. Int J Qual Health Care. 2007 Dec;19(6):349–57. doi: http:// dx .doi .org/ 10 .1093/ intqhc/ mzm042 PMID: 17872937
35. Cutts FT, Ferrari MJ, Krause LK, Tatem AJ, Mosser JF. Vaccination strategies for measles control and elimination: time to strengthen local initiatives. BMC Med. 2021 Jan 5;19(1):2. doi: http:// dx .doi .org/ 10 .1186/ s12916 -020 -01843 -z PMID: 33397366
36. Adamu AA, Jalo RI, Habonimana D, Wiysonge CS. COVID-19 and routine childhood immunization in Africa: leveraging systems thinking and implementation science to improve immunization system performance. Int J Infect Dis. 2020 Sep;98:161–5. doi: http:// dx .doi .org/ 10 .1016/ j .ijid .2020 .06 .072 PMID: 32592908
37. Leader AE, Burke-Garcia A, Massey PM, Roark JB. Understanding the messages and motivation of vaccine hesitant or refusing social media influencers. Vaccine. 2021 Jan 8;39(2):350–6. doi: http:// dx .doi .org/ 10 .1016/ j .vaccine .2020 .11 .058 PMID: 33280856
38. The Confidence Project [internet]. London: Vaccine Confidence Project; 2021. Available from: https:// www .vaccineconfidence .org/ [cited 2021 Sep 7].
39. Lane S, MacDonald NE, Marti M, Dumolard L. Vaccine hesitancy around the globe: analysis of three years of WHO/UNICEF Joint Reporting Form data 2015–2017. Vaccine. 2018 Jun 18;36(26):3861–7. doi: http:// dx .doi .org/ 10 .1016/ j .vaccine .2018 .03 .063 PMID: 29605516
40. Larson HJ, Schulz WS, Tucker JD, Smith DMD. Measuring vaccine confidence: introducing a global vaccine confidence index. PLoS Curr. 2015 Feb 25;7:ecurrents.outbreaks.ce0f6177bc97332602a8e3fe7d7f7cc4. doi: http:// dx .doi .org/ 10 .1371/ currents .outbreaks .ce 0f6177bc97 332602a8e3 fe7d7f7cc4 PMID: 25789200
Bull World Health Organ 2021;99:783–794D| doi: http://dx.doi.org/10.2471/BLT.20.285258 794A
ResearchVaccine hesitancy mechanisms, RwandaCatherine Decouttere et al.
Tabl
e 2.
Va
ccin
atio
n se
rvice
pro
vide
rs’ c
omm
ents
on
fact
ors a
ffect
ing
vacc
ine
hesit
ancy
, Rw
anda
, 201
8–20
20
Fact
or a
ffect
ing
vacc
ine
hesit
ancy
a
Sele
cted
com
men
ts o
n fa
ctor
by s
ervi
ce p
rovi
ders
b
Loca
tion
of se
rvice
pro
vide
rs’ f
acili
tiesc,d
Dist
rict h
ospi
tals
(n =
3)Ru
ral h
ealth
cent
res (
n =
5)Pe
ri-ur
ban
heal
th ce
ntre
s (n
= 6)
Confi
denc
eTr
ust i
n th
e eff
ectiv
enes
s an
d sa
fety
of
vacc
ines
and
in th
eir
man
ufac
ture
rs
Posit
ive:
“vac
cine
hes
itanc
y, in
the
narro
wes
t se
nse
of tr
ust i
n th
e va
ccin
e, is
not
an
issue
in
Rwan
da”
Posit
ive:
(i) h
igh
leve
l of t
rust
in v
acci
natio
n; a
nd (i
i) on
e va
ccin
ator
saw
ver
y fe
w a
dole
scen
t girl
s who
fear
ed th
e H
PV v
acci
ne a
nd h
ad q
uest
ions
abo
ut
rum
ours
that
the
vacc
ine
wou
ld p
reve
nt p
regn
ancy
Posit
ive:
hig
h le
vel o
f tru
st in
vac
cina
tion
Trus
t in,
and
per
sona
l ex
perie
nce
of, t
he h
ealth
sy
stem
and
hea
lth
prof
essio
nals
Neu
tral:
CHW
s are
hig
hly
resp
ecte
d (m
ore
in ru
ral
than
urb
an a
reas
). N
egat
ive:
(i) h
igh
turn
over
of s
taff
at h
ealth
ce
ntre
s; (ii
) ins
uffici
ent t
rain
ing
of v
acci
nato
rs
to u
pgra
de th
eir k
now
ledg
e; (i
ii) in
suffi
cien
t in
duct
ion
for n
ewly
recr
uite
d nu
rses
and
va
ccin
ator
s; (iv
) sho
rtag
e of
vac
cina
tors
and
nu
rses
; and
(v) h
igh
vacc
inat
or w
orkl
oad
due
to
mon
itorin
g an
d pa
perw
ork
Neu
tral:
som
e ca
regi
vers
trav
el fa
r to
see
an e
xper
ienc
ed n
urse
they
kno
w
rath
er th
an a
ttend
a n
ew, c
lose
r hea
lth c
entre
. N
egat
ive:
(i) s
hort
age
of st
aff a
t hea
lth c
entre
s bec
ause
nur
ses p
refe
r to
get
jobs
at a
hos
pita
l or i
n ci
ties r
athe
r tha
n in
rura
l are
as; (
ii) in
suffi
cien
t tra
inin
g of
nur
ses i
n va
ccin
e m
anag
emen
t; an
d (ii
i) du
ring
lock
dow
n, p
eopl
e w
ere
afra
id o
f get
ting
infe
cted
with
SAR
S-Co
V-2
whe
n th
ey v
isite
d he
alth
cen
tres
Posit
ive:
con
nect
ions
with
hea
lth c
entre
s mad
e du
ring
ante
nata
l car
e an
d in
stitu
tiona
l birt
hs.
Neu
tral:
the
impa
ct o
f loc
kdow
n on
hea
lth c
entre
vi
sits w
as h
ighe
r in
rura
l are
as.
Neg
ativ
e: (i
) ins
uffici
ent t
ime
with
car
egiv
ers
beca
use
of p
aper
wor
k lo
ad; (
ii) to
o m
uch
wor
k fo
r th
e st
aff a
vaila
ble;
and
(iii)
staff
shor
tage
s a h
uge
prob
lem
Com
plac
ency
Com
mun
icat
ion
and
med
ia e
nviro
nmen
tPo
sitiv
e: m
othe
r-an
d-ch
ild h
ealth
wee
ks h
eld
ever
y 6
mon
ths
Posit
ive:
CH
Ws’
role
incr
ease
d du
ring
lock
dow
n (w
hen
they
wer
e th
e on
ly
chan
nel f
or in
form
atio
n)N
one
Influ
entia
l lea
ders
, im
mun
izat
ion
prog
ram
me
gate
keep
ers
and
vacc
inat
ion
lobb
ies
Posit
ive:
hea
lth m
inist
er su
ppor
tive
of
imm
uniz
atio
n pr
ogra
mm
e, C
HW
s and
mot
her-
and-
child
hea
lth w
eeks
Non
eN
one
Relig
ion,
cul
ture
, gen
der
and
soci
oeco
nom
ic
fact
ors
Non
ePo
sitiv
e: u
mug
anda
(i.e
. mon
thly
Rw
anda
n vo
lunt
ary
com
mun
ity m
eetin
gs
whe
re lo
cal i
ssue
s and
upd
ates
are
disc
usse
d).
Neg
ativ
e: (i
) um
ugan
da w
as te
mpo
raril
y su
spen
ded
durin
g lo
ckdo
wn;
(ii
) “m
othe
rs a
re fo
rget
ting
abou
t the
seco
nd m
easle
s dos
e”; (
iii) “
atte
ntio
n of
m
othe
r red
uces
whe
n ch
ild g
row
s up”
; (iii
) mot
hers
hav
e ot
her p
riorit
ies a
nd
post
pone
visi
ts to
hea
lth c
entre
s and
vac
cina
tion
sess
ions
, res
ultin
g in
late
im
mun
izat
ion;
(iv)
“for
gett
ing
is no
t abo
ut b
eing
unw
illin
g, b
ut th
e m
othe
r co
uldn
’t go
bec
ause
of o
ther
prio
ritie
s”; (
v) “I
wou
ld g
o to
mor
row
or n
ext
wee
k”; (
vi) s
omet
imes
the
vacc
inat
ion
sess
ion
is th
e sa
me
day
whe
n sa
larie
s fo
r 15
days
are
giv
en o
ut, s
o ca
regi
vers
can
not a
ttend
that
day
; and
(vii)
poo
r fa
mili
es w
ork
long
hou
rs a
way
from
hom
e, le
avin
g lit
tle ti
me
for c
hild
ren
Non
e
(contin
ues.
. .)
Bull World Health Organ 2021;99:783–794D| doi: http://dx.doi.org/10.2471/BLT.20.285258794B
ResearchVaccine hesitancy mechanisms, Rwanda Catherine Decouttere et al.
Fact
or a
ffect
ing
vacc
ine
hesit
ancy
a
Sele
cted
com
men
ts o
n fa
ctor
by s
ervi
ce p
rovi
ders
b
Loca
tion
of se
rvice
pro
vide
rs’ f
acili
tiesc,d
Dist
rict h
ospi
tals
(n =
3)Ru
ral h
ealth
cent
res (
n =
5)Pe
ri-ur
ban
heal
th ce
ntre
s (n
= 6)
Know
ledg
e an
d aw
aren
ess
Posit
ive:
(i) C
HW
s hav
e an
impo
rtan
t rol
e in
bu
ildin
g co
mm
unity
eng
agem
ent;
(ii) C
HW
s are
tru
sted
and
hig
hly
resp
ecte
d by
the
com
mun
ity;
and
(iii)
hom
e vi
sits b
y CH
Ws a
t chi
ldbi
rth
conn
ects
mot
hers
with
hea
lth c
entre
s and
in
form
s the
m a
bout
imm
uniz
atio
n.
Neg
ativ
e: C
HW
serv
ice
may
not
be
sust
aina
ble
beca
use
repl
acin
g ag
eing
CH
Ws d
epen
ds o
n th
e m
otiv
atio
n of
the
next
gen
erat
ion
Posit
ive:
(i) 4
5- to
60-
min
ute
info
rmat
ion
sess
ions
bef
ore
vacc
inat
ion
sess
ions
at
hea
lth c
entre
s and
out
reac
h po
sts;
and
(ii) v
ery
low
per
cent
age
of p
eopl
e ha
ve li
ttle
und
erst
andi
ng o
f the
impo
rtan
ce o
f vac
cine
s. N
egat
ive:
(i) i
nfor
mat
ion
sess
ions
at f
acili
ties w
ere
susp
ende
d du
ring
lock
dow
n if
ther
e w
as in
suffi
cien
t spa
ce; (
ii) th
ere
was
a lo
ng in
terv
al w
ith
no c
onta
ct w
ith h
ealth
cen
tres b
etw
een
child
hood
vac
cina
tions
at 9
and
15
mon
ths o
f age
; and
(iii)
dur
ing
the
2019
mea
sles o
utbr
eak,
CH
Ws w
ere
not
able
to v
isit t
he w
hole
are
a to
raise
aw
aren
ess a
nd d
etec
t mea
sles c
ases
Non
e
Perc
eive
d ris
ks a
nd
bene
fits
Non
eN
egat
ive:
(i) t
he p
erce
ived
risk
of d
iseas
e w
as lo
w b
ecau
se c
ases
in th
e co
mm
unity
rem
aine
d un
dete
cted
; and
(ii)
“mea
sles c
ases
wer
e no
t det
ecte
d,
not d
iagn
osed
, and
not
med
ical
ly tr
eate
d”
Posit
ive:
“peo
ple
are
intri
nsic
ally
mot
ivat
ed fo
r va
ccin
atio
n”
Imm
uniz
atio
n as
a so
cial
no
rmPo
sitiv
e: im
mun
izat
ion
disc
usse
d as
par
t of
mon
thly
com
mun
ity m
eetin
gs (u
mug
anda
) and
m
othe
rs’ e
veni
ngs (
w'a
baby
eyi)
Non
eN
one
Conv
enie
nce
Avai
labi
lity
of th
e im
mun
izat
ion
serv
ice
Posit
ive:
CH
Ws’
role
in ra
ising
aw
aren
ess i
n th
e co
mm
unity
and
org
aniz
ing
outre
ach
serv
ices
. N
egat
ive:
hig
her v
acci
ne w
asta
ge a
t out
reac
h po
sts
Neg
ativ
e: (i
) poo
r fam
ilies
wer
e no
t abl
e to
affo
rd tr
avel
to h
ealth
cen
tres a
nd
relie
d on
out
reac
h se
rvic
es; (
ii) m
easle
s and
rube
lla v
acci
ne a
vaila
ble
only
at
heal
th c
entre
s and
not
at o
utre
ach
post
s; (ii
i) BC
G va
ccin
e av
aila
ble
only
whe
n 10
chi
ldre
n w
ere
wai
ting;
(iv)
mot
hers
that
gav
e bi
rth
som
etim
es n
eede
d to
com
e ba
ck to
the
heal
th c
entre
with
in 2
wee
ks fo
r the
BCG
vac
cine
; and
(v
) vac
cine
s in
mul
tidos
e vi
als (
e.g.
the
mea
sles a
nd ru
bella
vac
cine
) wer
e off
ered
onc
e a
wee
k w
here
as o
ther
vac
cine
s wer
e off
ered
dai
ly
Posit
ive:
(i) o
utre
ach
serv
ices
wer
e off
ered
at t
wo
sites
(2 d
ays p
er w
eek
at o
ne a
nd 1
day
per
wee
k at
th
e ot
her);
(ii)
CHW
s hel
ped
incr
ease
the
effici
ency
of
out
reac
h se
rvic
es b
y m
akin
g su
re th
e rig
ht
amou
nt o
f vac
cine
was
take
n to
out
reac
h po
sts;
and
(iii)
all v
acci
nes,
incl
udin
g th
e BC
G va
ccin
e, o
ffere
d ev
ery
day.
Neg
ativ
e: v
acci
nes i
n m
ultid
ose
vial
s (e.
g. th
e m
easle
s and
rube
lla v
acci
ne) w
ere
offer
ed o
nly
once
a
wee
k w
here
as o
ther
vac
cine
s wer
e off
ered
dai
lyAff
orda
bilit
y of
the
imm
uniz
atio
n se
rvic
eN
one
Neg
ativ
e: p
oor f
amili
es w
ere
not a
ble
to a
fford
to tr
avel
to h
ealth
cen
tres
Posit
ive:
“pov
erty
is n
ot a
reas
on fo
r not
com
ing,
ev
eryb
ody
com
es: p
oor a
nd le
ss p
oor.”
N
egat
ive:
“usu
ally
, liv
ing
cond
ition
s are
a d
river
for
drop
ping
out
”G
eogr
aphi
cal a
cces
sibili
tyN
one
Posit
ive:
“sho
rt d
istan
ce, p
eopl
e fro
m d
iffer
ent c
atch
men
t are
as a
nd d
istric
ts
com
e he
re.”
Neg
ativ
e: (i
) hea
lth c
entre
diffi
cult
to a
cces
s; an
d (ii
) hill
y ar
ea w
ith lo
ng
dist
ance
s to
heal
th c
entre
s, so
peo
ple
rely
on
outre
ach
Neg
ativ
e: re
stric
ted
acce
ss to
one
hea
lth c
entre
due
to
a la
ndsli
de
Abili
ty to
und
erst
and
(i.e.
lang
uage
and
hea
lth
liter
acy)
Neg
ativ
e: m
othe
rs so
met
imes
forg
et th
e sc
hedu
led
appo
intm
ent f
or th
e ne
xt v
acci
nes
Neg
ativ
e: (i
) con
fusio
n ab
out t
he se
cond
mea
sles a
nd ru
bella
vac
cine
dos
e fo
r ch
ildre
n ag
ed 1
5 m
onth
s bec
ause
mos
quito
net
s are
disp
ense
d w
hen
they
ar
e ag
ed 9
mon
ths;
(ii) m
othe
rs’ la
ck o
f edu
catio
n is
a re
ason
for t
he se
cond
m
easle
s and
rube
lla v
acci
ne d
ose
bein
g m
issed
; and
(iii)
som
e ca
regi
vers
are
no
t abl
e to
read
the
vacc
inat
ion
card
bec
ause
of a
lack
of e
duca
tion
Posit
ive:
app
oint
men
t sys
tem
with
car
ds is
wel
l un
ders
tood
by
care
give
rs.
Neg
ativ
e: m
isund
erst
andi
ng o
f pra
ctic
al a
nd
orga
niza
tiona
l pol
icie
s mak
es p
eopl
e fe
el
unw
elco
me
(. . .continued)
(contin
ues.
. .)
Bull World Health Organ 2021;99:783–794D| doi: http://dx.doi.org/10.2471/BLT.20.285258 794C
ResearchVaccine hesitancy mechanisms, RwandaCatherine Decouttere et al.
Fact
or a
ffect
ing
vacc
ine
hesit
ancy
a
Sele
cted
com
men
ts o
n fa
ctor
by s
ervi
ce p
rovi
ders
b
Loca
tion
of se
rvice
pro
vide
rs’ f
acili
tiesc,d
Dist
rict h
ospi
tals
(n =
3)Ru
ral h
ealth
cent
res (
n =
5)Pe
ri-ur
ban
heal
th ce
ntre
s (n
= 6)
Qua
lity
of th
e se
rvic
e (p
erce
ived
or r
eal)
Non
ePo
sitiv
e: (i
) CH
Ws c
lose
ly in
volv
ed in
out
reac
h or
gani
zatio
n an
d in
trac
ing
vacc
ine
defa
ulte
rs; a
nd (i
i) fe
wer
pat
ient
s atte
nded
hea
lth c
entre
s dur
ing
lock
dow
n, le
avin
g m
ore
time
for e
ach
patie
nt.
Neg
ativ
e: (i
) hig
h nu
rse
wor
kloa
d du
e to
com
bini
ng v
acci
natio
n se
ssio
ns a
t he
alth
cen
tres a
nd o
utre
ach
post
s with
nig
ht sh
ifts;
(ii) i
ncre
ased
num
ber
of c
hild
ren
to b
e va
ccin
ated
in e
ach
sess
ion;
(iii)
one
hea
lth c
entre
did
no
t fun
ctio
n w
ell i
n th
e pa
st a
s it p
rovi
ded
only
fixe
d se
ssio
ns a
nd h
ad n
o ou
treac
h fo
r sev
eral
yea
rs; a
nd (i
v) 7
3 ch
ildre
n m
issed
the
seco
nd m
easle
s and
ru
bella
vac
cine
dos
e in
201
8
Posit
ive:
“spl
ittin
g up
the
larg
e ca
tchm
ent a
rea
led
to b
ette
r man
agem
ent.
All v
acci
nes a
re n
ow o
ffere
d ev
ery
day,
exce
pt th
e BC
G va
ccin
e.”
Neg
ativ
e: (i
) “so
met
imes
the
EPI n
urse
has
a n
ight
sh
ift th
e da
y be
fore
the
sess
ion.
Som
etim
es sh
e ca
nnot
atte
nd th
e se
ssio
n”; (
ii) th
e po
pula
tion
is in
crea
sing
at 1
0–15
% p
er a
nnum
but
serv
ices
ca
nnot
kee
p up
; and
(iii)
exp
erie
nced
mot
hers
saw
a
drop
in se
rvic
e qu
ality
due
to in
crea
sed
nurs
e w
orkl
oad
as m
ore
prog
ram
mes
wer
e de
cent
raliz
ed
to h
ealth
cen
tres a
nd h
ealth
pos
ts (e
.g. m
alar
ia,
tube
rcul
osis
and
HIV
serv
ices
and
fam
ily p
lann
ing)
Conv
enie
nt ti
me
(incl
udin
g w
aitin
g tim
e),
plac
e an
d cu
ltura
l con
text
Neg
ativ
e: (i
) miss
ed o
ppor
tuni
ties t
o va
ccin
ate
beca
use
vacc
inat
ion
days
wer
e di
ffere
nt in
di
ffere
nt h
ealth
cen
tres;
and
(ii) C
OVI
D-1
9 re
stric
tions
resu
lted
in a
lack
of i
ndoo
r wai
ting
room
s, m
akin
g it
hard
for c
areg
iver
s and
chi
ldre
n
Posit
ive:
(i) p
rovi
ding
nut
ritio
n se
rvic
es a
nd fa
mily
pla
nnin
g in
add
ition
to
vacc
inat
ion
was
rece
ived
pos
itive
ly; a
nd (i
i) ha
ving
vac
cina
tion
sess
ions
on
mar
ket d
ays e
ncou
rage
s mot
hers
to a
ttend
. N
eutra
l: va
ccin
atio
n is
the
mot
her’s
resp
onsib
ility
and
she
pref
ers t
o co
me
to
the
heal
th c
entre
whe
n th
ere
is a
mar
ket n
earb
y. N
egat
ive:
pro
blem
with
pow
er b
acku
p at
a h
ealth
cen
tre
Posit
ive:
pro
vidi
ng n
utrit
ion
serv
ices
in a
dditi
on to
va
ccin
atio
n.
Neu
tral:
the
day
of th
e w
eek
is le
ss im
port
ant t
han
whe
ther
it is
rain
ing.
N
egat
ive:
(i) w
ith 4
0–50
peo
ple
per s
essio
n, w
aitin
g tim
es c
an b
e up
to 5
hou
rs; (
ii) 6
0–75
or e
ven
100
child
ren
per s
essio
n; a
nd (i
ii) n
urse
wan
ts to
giv
e BC
G va
ccin
e an
d m
easle
s and
rube
lla v
acci
ne o
n th
e sa
me
day
but B
CG v
acci
nes t
ake
time
as th
ey re
quire
a
new
file
and
car
d to
be
mad
e ea
ch ti
me
Des
ign
of v
acci
natio
n pr
ogra
mm
e, v
acci
natio
n sc
hedu
le a
nd d
ata
man
agem
ent a
t hea
lth
cent
res
Posit
ive:
the
trans
ition
to a
dig
ital d
ata
man
agem
ent s
yste
m w
ill im
prov
e th
e fu
nctio
ning
of v
acci
natio
n se
ssio
ns
Neg
ativ
e: (i
) “de
faul
ter t
raci
ng d
idn’
t wor
k w
ell [
due
to p
aper
-bas
ed d
ata
syst
em a
nd h
igh
wor
kloa
d] a
nd p
eopl
e ar
e lo
st to
follo
w-u
p w
hen
they
go
to
the
neig
hbou
ring
dist
rict”
; and
(ii)
data
man
agem
ent s
yste
m d
oes n
ot sh
ow
whe
n pe
ople
are
dro
ppin
g ou
t
Posit
ive:
“Com
pute
rs w
ill b
e in
stal
led
soon
.” N
egat
ive:
(i) i
dent
ifyin
g dr
op-o
uts u
sing
pape
r file
s is
labo
ur-in
tens
ive
and
requ
ires t
he h
elp
of C
HW
s; (ii
) “dr
op-o
uts a
re c
heck
ed a
t the
end
of t
he m
onth
, re
ason
s for
dro
ppin
g ou
t are
exp
lore
d. T
hey
are
usua
lly li
ving
con
ditio
ns, o
r whe
n pe
ople
mov
e to
ano
ther
are
a”; (
iii) c
alcu
latin
g va
ccin
e co
vera
ge
is di
fficu
lt be
caus
e of
the
chan
ging
pop
ulat
ion
in
the
catc
hmen
t are
a; (i
v) a
t the
age
of 9
mon
ths,
child
ren
need
to g
et th
e fir
st m
easle
s and
rube
lla
vacc
ine
dose
at t
heir
own
heal
th c
entre
to re
ceiv
e a
mos
quito
net
; and
(v) “
no fr
ee c
hoic
e of
hea
lth
cent
re fo
r mea
sles a
nd ru
bella
”
BCG:
bac
illus
Cal
met
te–G
uérin
; CHW
: com
mun
ity h
ealth
wor
ker;
COVI
D-1
9: c
oron
aviru
s dise
ase
2019
; EPI
: Exp
ande
d Pr
ogra
mm
e on
Imm
uniza
tion;
HIV
: hum
an im
mun
odefi
cien
cy v
irus;
HPV:
hum
an p
apill
omav
irus;
SARS
-CoV
-2: s
ever
e ac
ute
resp
irato
ry sy
ndro
me
coro
navi
rus 2
.a F
acto
rs w
ere
cate
goriz
ed u
sing
the
3Cs f
ram
ewor
k as
rela
ted
to c
onfid
ence
, com
plac
ency
or c
onve
nien
ce.8
b Com
men
ts w
ere
clas
sified
as p
ositi
ve, n
eutra
l or n
egat
ive
with
rega
rd to
thei
r im
plic
atio
ns fo
r vac
cine
upt
ake.
c Vac
cina
tors
and
Exp
ande
d Pr
ogra
mm
e on
Imm
uniza
tion
staff
wer
e in
terv
iew
ed.
d Det
ails
of w
hich
inte
rvie
wee
s mad
e ea
ch c
omm
ent a
re a
vaila
ble
from
supp
lem
ent 3
in th
e da
ta re
posit
ory.32
(. . .continued)
Bull World Health Organ 2021;99:783–794D| doi: http://dx.doi.org/10.2471/BLT.20.285258794D
ResearchVaccine hesitancy mechanisms, Rwanda Catherine Decouttere et al.
Fig. 2. Causal loop diagrams of factors affecting vaccine hesitancy, by factor category, Rwanda, 2018–2020
Pressure on immunization
system
Disease incidence
–
+
Protected population
R: Reinforcing loop B: Balancing loop
+
+
++
+
+
Vaccination
Confidence Convenience
Community engagement
Perceived vaccine benefit
Vaccination–
+ Positive effect – Negative effect
Complacency
–
+
++
–
Relationship with health centre staff
Perceived quality of vaccination service
Trust in vaccination
Perceived threat of disease
R Complacency
B Complacency
R Confidence
BConfidence
Vaccination
Outreach and campaign planning
Outreach and campaign need
Individual health
R Convenience
B Convenience
Access to vaccination
Health literacy and mobility
Protected population
+
+
++
+
+
+
Notes: Factors affecting vaccine hesitancy were categorized using the 3Cs framework as related to confidence, complacency or convenience.8 The R loops are reinforcing feedback loops that cause change in the same direction and the B loops are balancing loops that cause change in the opposite direction.For confidence, reinforcing factors represented by the R loop include: (i) respectful relationships with health centre staff and community health workers engenders trust in vaccine services; (ii) mothers overcome fear of being late for immunization when they know the nurse and feel welcome; (iii) multiple contacts between the caregiver and nurse increases trust; and (iv) an entry in the mother-and-child booklet indicating immunization has been completed is encouraging for both caregivers and their families. Balancing factors represented by the B loop include: (i) long waiting times; (ii) short contact times between nurse and caregiver, which give the nurse little time to explain vaccination (one reason was numerous administrative tasks and demands from other vaccination programmes); and (iii) suspension of educational sessions before vaccinations during the coronavirus disease 2019 pandemic.For complacency, reinforcing factors represented by the R loop include: (i) a well-protected population experiences fewer cases of disease, which increases the perceived benefit of vaccination and community engagement. Balancing factors represented by the B loop include: (i) a decline in cases of disease lowers both awareness of the threat to young children and the perceived importance of vaccination; and (ii) caregivers worry less about their children’s health as they grow up, which increases complacency and can lead some to deprioritize their vaccination appointments.For convenience: reinforcing factors represented by the R loop include: (i) families that are healthy, literate and financially stable are more likely to understand their appointment schedule and can afford to travel to health centres to vaccinate their children. Balancing factors represented by the B loop include: (i) the risk of a disease outbreak is increased in poorly protected populations (which can be remedied by immunization campaigns and enhanced vaccination services, including outreach).