Implementation research to support Bangladesh Ministry of ... · F or 35.5 ° C † Raised temperature (> 99.5 ° F or 37.5 ° C) † Less movement/movement only when stimulated †

STUDY PROTOCOL Open Access

Implementation research to supportBangladesh Ministry of Health and FamilyWelfare to implement its nationalguidelines for management of infections inyoung infants in two rural districtsSalahuddin Ahmed1, Jennifer A. Applegate2, Dipak K. Mitra3, Jennifer A. Callaghan-Koru4, Mahfuza Mousumi5,Ahad Mahmud Khan1, Taufique Joarder6, Meagan Harrison2, Sabbir Ahmed7, Nazma Begum1, Abdul Quaiyum8,Joby George7 and Abdullah H. Baqui2*

Abstract

Background: World Health Organization revised the global guidelines for management of possible serious bacterialinfection (PSBI) in young infants to recommend the use of simplified antibiotic therapy in settings where access tohospital care is not possible. The Bangladesh Ministry of Health and Family Welfare (MoHFW), Government ofBangladesh (GOB) adopted these guidelines, allowing treatment at first-level facilities. During the first year ofimplementation, the Projahnmo Study Group and USAID/MaMoni Health Systems Strengthening (HSS) Projectsupported the MoHFW to operationalize the new guidelines and conducted an implementation research study inselected districts to assess challenges and identify solutions to facilitate scale-up across the country.

Implementation support: Projahnmo and MaMoni HSS teams supported implementation in three areas: buildingcapacity, strengthening service delivery, and mobilizing communities. Capacity building focused on trainingparamedics to conduct outpatient management of PSBI cases and developing monitoring and supervision systems.The teams also filled gaps in government supply of essential drugs, equipment, and logistics. Communitymobilization strategies to promote care-seeking and referrals to facilities varied across districts; in one districtcommunity, health workers made home visits while in another district, the promotion was carried out throughcommunity volunteers, village doctors, and through existing community structures.

Methods: We followed a plan-do-study-act (PDSA) cycle to identify and address implementation challenges. Threecycles—1 every 4 months—were conducted. We collected data utilizing quantitative and qualitative methods inboth the community and facilities. The total sample size for this study was 13,590.

(Continued on next page)

© The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

* Correspondence: [email protected] Center for Maternal and Newborn Health, Department ofInternational Health, Johns Hopkins Bloomberg School of Public Health,Baltimore, MD 21205, USAFull list of author information is available at the end of the article

Ahmed et al. Journal of Health, Population and Nutrition (2019) 38:41 https://doi.org/10.1186/s41043-019-0200-6

http://crossmark.crossref.org/dialog/?doi=10.1186/s41043-019-0200-6&domain=pdfhttp://creativecommons.org/licenses/by/4.0/http://creativecommons.org/publicdomain/zero/1.0/mailto:[email protected]

(Continued from previous page)

Discussion: This article provides implementation research design details for program managers intending toimplement new guidelines on management of young infant infections. Results of this research will be reported inthe forthcoming papers. Preliminary findings indicate that the management of PSBI cases at the UH&FWCs isfeasible. However, MoHFW, GOB needs to address the implementation challenges before scale-up of this policy tothe national level.

Keywords: Implementation research, Young infant infection, Possible serious bacterial infection, Outpatientmanagement; Bangladesh

BackgroundGlobal rates of child mortality have dropped significantlyover the past few decades, with remarkable declines seenfor older children. However, mortality rates in neonatesremain high with an estimated 2.7 million annual deathsglobally [1]. About 45% of all deaths in children under 5years of age occur during the neonatal period [2] and al-most 98% of neonatal deaths occur in developing coun-tries [1]. Globally, an estimated one-fourth of theneonatal deaths are attributed to infectious causes, andin settings characterized by high neonatal mortalityrates, the proportion of neonatal deaths due to infectionsis estimated to be even higher [3, 4]. In Bangladesh,newborn infections remain a major cause of both mor-bidity and mortality [4, 5]. About 37% of all neonataldeaths in Bangladesh occur as a result of sepsis or othersevere infections [6].The World Health Organization (WHO) recom-

mends that young infants (0-59 days) with signs ofpossible serious bacterial infection (PSBI) be referredto hospitals for treatment with a 7–10 day course oftwo injectable antibiotics–penicillin (or ampicillin)and gentamicin. However, referral compliance forhospitalization in many developing countries has beenlow due to limited access or inadequate hospital facil-ities [7, 8]. In 2007, the WHO, United States Agencyfor International Development (USAID), and Save theChildren’s Saving Newborn Lives program (SC/SNL)convened an expert panel aimed at identifying simple,safe, and effective treatment regimens that could beprovided to young infants with severe infectionscloser to home when the family was not able toaccept referral to the hospital [9]. The panel con-cluded that the existing evidence was insufficient torecommend antibiotic treatment for severe infectionsat the community level and identified the need foradditional research on the efficacy of simplified anti-biotic therapy [9]. Three randomized, open-label,equivalence trials were conducted in Bangladesh,Pakistan and three countries in Africa (DemocraticRepublic of Congo [DRC], Kenya, and Nigeria) toevaluate the efficacy of simplified antibiotic regimensfor managing PSBI in young infants at the community

level when referral was not possible [10]. While thetrial protocols were harmonized, the number of dosesand service delivery mechanisms varied across thestudies. Findings from all three studies demonstratedthat the simplified regimens were as efficacious as thestandard regimen [11–13].In 2015, the WHO revised the global guidelines

recommending use of simplified antibiotic regimens forthe management of PSBI in young infants for resource-limited settings when hospitalization is not acceptable oraccessible to families [14]. The Government ofBangladesh (GOB) adopted the WHO guidelines and de-veloped a corresponding policy, titled Management ofInfection of the 0–59 Days Infants at Union Level Facil-ities and NGO Clinics without Indoor Facilities [15]. Theunion level facilities under the management of the Min-istry of Health and Family Welfare (MoHFW) in ruralBangladesh are known as health and family welfare cen-ters (UH&FWCs). In most administrative unions, thereis one UH&FWC, which serves a catchment populationof approximately 25,000 persons [16, 17]. TheUH&FWC provides mostly outpatient services. The ser-vices offered at the UH&FWCs include essential mater-nal, newborn, child health, family planning, andnutrition services, including the management of normalvaginal deliveries. It is staffed with one Sub-AssistantCommunity Medical Officer (SACMO) who has at least3 years of training on general health care including childhealth and at least one Family Welfare Visitor (FWV)who has at least 18 months of training on pregnancycare and family planning.The Comprehensive Newborn Care Package

(CNCP) was developed for the implementation ofnewly recommended priority newborn interventions,including management of infections in young in-fants. With implementation of the new guidelines,SACMOs are being trained with CNCP to assessand treat infants with PSBI. Per the updated guide-lines, the SACMO assesses the infant and deter-mines an illness classification based on thestandardized Integrated Management of ChildhoodIllness (IMCI) algorithm for infants under 2 monthsof age (Table 1).

Ahmed et al. Journal of Health, Population and Nutrition (2019) 38:41 Page 2 of 13

Table

1Ope

ratio

nalalgorith

mformanaginginfections

inyoun

ginfantsin

UH&F

WCpe

rtheBang

lade

shgu

idelines

Categ

ory

Clinicalsign

sManagem

ent

Follow-upandreferralsupp

ort

Criticalillne

sses

(CI)

•Uncon

scious/drowsy

•Con

vulsion/historyof

convulsion

•Unableto

feed

•Persistent

vomiting

•Cen

tralcyanosis

•Bu

lgingfontanel

•Weigh

t<1500

g

•Theyoun

ginfant

willbe

administeredthe

1stdo

seof

injectablege

ntam

icin

andoral

antib

iotic

(ifpo

ssible),advisedabou

tthe

impo

rtance

ofho

spitalizationandreferred

urge

ntlyto

thede

sign

ated

referralfacility

with

areferralslip

containing

referralno

tes

ofSA

CMO

•Themothe

rwillbe

advisedon

frequ

ent

breastfeed

ingto

preven

tlow

bloo

dsugar.

Shewillalso

beprop

erlyadvisedto

keep

thebaby

warm

espe

ciallydu

ring

transportatio

n.

•Themob

ileph

onecontactnu

mbe

rwillbe

kept

tofollow-upthereferralcompliance.Theph

one

numbe

rof

theSA

CMOwillbe

provided

tothefamily

•TheUH&F

WCserviceproviderswillcommun

icatewith

theUpazilaHealth

Com

plex

(UHC)(re

ferralcenter)

abou

tthecase

•Necessary

supp

ortto

beprovided

byUH&F

WCservice

providersor

field

supe

rvisorsto

arrang

etransportfor

referral

Clinicalsevere

infection(CSI)*

•Severe

chestin-drawing

•Hypothe

rmia(<

95.9

° For

35.5

° C•Raised

tempe

rature

(>99.5

° For

37.5

° C)

•Less

movem

ent/movem

enton

lywhe

nstim

ulated

•Not

feed

ingwell(de

pend

ingon

historyand

observation)

•Thecase

willbe

administered1stdo

seinjectablege

ntam

icin

andoralam

oxicillin,

andreferred

followingtheaboveproced

ure

tothene

arestUHCformanagem

ent

•Sameas

above,theSA

CMO’smob

ilenu

mbe

rwillbe

givento

caregiverandthecase

willbe

followed

upover

phon

eto

record

referralcomplianceon

theday

ofreferralby

UH&F

WCprovider

Incase

ofreferralno

n-compliance:

•Theinfant

willbe

managed

bytheSA

CMO

usingstandard

managem

entprotocol:

oInjectionge

ntam

icin

I/Mon

cedaily

atUH&F

WCfor2days

oOralamoxicillin

twicedaily

for7days

•Thefamily

willbe

coun

seledand

advisedto

cometo

thesamefacility

with

thebaby

toreceivethe2n

d(last)

dose

ofinjectableantib

iotic

andcontinue

oralmed

icine12

hourlyfortotal7

days

•Onthe2n

ddayof

treatm

ent,theinfant

shou

ldreturn

toUH&F

WCforassessmen

tand2n

ddo

seinjectablege

ntam

icin

•Onthe4thand8thdayof

treatm

ent,follow-up

willbe

cond

uctedto

assess

cond

ition

oftheinfant

•Ifthebaby

develops

anyne

wsymptom

(listed

symptom

sof

CSIor

CI),

does

notim

proveafter4days

ofreceivingt

reatmen

tor,

isno

tfully

curedaftertreatm

entcompletion

(onthe8thday);the

family

shou

ldbe

advised

forim

med

iate

notificationto

thesameservice

provider

andto

seek

care

from

referralfacility

Isolated

fast-breathing

assing

lesign

ofillne

ss•Yo

unginfants0–6days

oldwith

fastbreathingas

the

onlysign

ofillne

ss*

•Give1stdo

seof

oralam

oxicillin

andrefer

toUHC

•Themob

ileph

onecontactnu

mbe

rwillbe

kept

tofollow-upthereferralcompliance.Theph

one

numbe

rof

theSA

CMOwillbe

provided

tothe

family

•TheUH&F

WCserviceproviderswillcommun

icate

with

theUHC(re

ferralcenter)abou

tthecase

Incase

ofreferralno

n-compliance:

•Theinfant

willbe

managed

bytheSA

CMO

usingstandard

managem

entprotocol:

oOralamoxicillin

(100

mg/kg/day

twice

daily)for7days

•Infant

willbe

followed

upon

the4thdayand

8thday

•Ifthebaby

develops

anyne

wsymptom

(listed

symptom

sof

CSIor

CI)or,d

oesno

tim

prove

after4days

ofreceivingtreatm

ent,or

isno

tfully

curedaftertreatm

entcompletion(on

8thday),the

family

shou

ldbe

advisedfor

immed

iate

notificationto

thesameservice

provider

andto

seek

care

from

referralfacility

•Yo

unginfants7–59

days

oldwith

fastbreathingas

the

onlysign

ofillne

ss•Noreferral,treated

with

oralam

oxicillin

(100

mg/kg/day

twicedaily)for7days

•Sick

infantswith

fast-breathing

(7–59days)will

befollowed

upon

the4thdayand8thday


Table

1Ope

ratio

nalalgorith

mformanaginginfections

inyoun

ginfantsin

UH&F

WCpe

rtheBang

lade

shgu

idelines

(Con

tinued)

Categ

ory

Clinicalsign

sManagem

ent

Follow-upandreferralsupp

ort

Localb

acterialinfectio

n•Umbilicalredn

ess

•Drainingpu

sfro

mum

bilicus

•Skin

pustule

•Noreferral,treated

with

oralam

oxicillin

(125

mgdaily

forbe

low

1-mon

thaged

infantsor

infantshaving

less

than

4kg

weigh

tand250mgforinfantsaged

betw

een1and2mon

ths)for5days

•Careg

iver

willbe

advisedto

seek

immed

iate

consultatio

nwith

UH&F

WCprovider

ifinfant

does

notim

prove,ne

wsymptom

sappe

ar,

orcond

ition

worsens

*PSB

Icases

eligible

forsimplified

antib

iotic

treatm

entwhe

nho

spita

lreferralisno

tfeasible

forfamilies


If the SACMO identifies any signs of PSBI, then theSACMO is trained to administer the first dose of inject-able and/or oral antibiotics and refer the infant to theUpazila (sub-district) Health Complex (UHC). If thefamily declines referral to the hospital, then theSACMO either reinforces referral or treats the infantdepending upon classification per the guidelines,which also includes providing medicine to be admin-istered at home by the caregiver. The ability of theSACMO to treat infants with non-critical illnesses onan outpatient basis is the primary change to the pre-viously established treatment protocol. According tothe protocol, the SACMO also follows-up PSBI casesat day 4 through phone or in the facility if the parentbrings the infant for a follow-up visit. During followup, the SACMO decides whether to continue treat-ment (if condition improved) or refer to the higherfacility for further management (if condition has notimproved or new symptoms developed).The other cadres of providers involved in outpatient

management of PSBI cases are the FWV and FamilyPlanning Inspectors (FPI). The FWVs are posted atthe UH&FWC and primarily provide antenatal care,normal delivery care, postnatal care and family plan-ning services to the community. FWVs are able toprovide the second dose of injectable gentamicin toPSBI cases in the absence of the Sub-Assistant Com-munity Medical Officers (SACMO) [15]. FPIs arenon-clinical field supervisors of frontline workers inthe community. For PSBI management, the FPIs aretrained and engaged for follow-up of the infant at theend of treatment (day 8 follow-up) within the com-munity. During these home visits, FPIs assess thecondition of the infant, record any existing signs orsymptoms, determine the condition of the infant (i.e.,recovered or not recovered), and advise on referral ifthe infant has not recovered.Prior to national scale-up of the guidelines, the

Bangladesh MoHFW planned to learn from imple-mentation of the policy in three selected districts ofBangladesh: Kushtia, Lakshmipur, and Sylhet. Weconducted an implementation research study in thefirst year of this program (September 2015–August2016) to document the inputs and processes requiredfor operationalization of the updated policy in varyingcontexts, identify barriers and facilitators for imple-mentation, and integrate these early lessons into theplans for national scale-up. This paper describes theimplementation research protocol followed by theProjahnmo and MaMoni Health System Strengthening(HSS) teams who provided support to the MoHFW inSylhet and Lakshmipur, respectively. A third partnerprovided support in Kushtia, but their methodology isnot described in this paper.

MethodsStudy settingThis implementation research was conducted in twosub-districts of Sylhet district in Sylhet division and onesub-district of Lakshmipur district in Chittagong division(Fig. 1). Sylhet and Chittagong are historically low per-forming divisions of Bangladesh for maternal, newborn,and child health indicators. According to the 2014Bangladesh Demographic and Health Survey, mothers inSylhet had the lowest proportion of births in facilities(22.6%) and lowest proportion of births attended by askilled provider (27.1%) [6], followed by Chittagong div-ision where 35.2% deliveries took place in facilities, and43.9% of the deliveries were attended by a medicallytrained provider [6].

Study designDuring the first year of implementation of theupdated PSBI guidelines, the MoHFW received imple-mentation support from Projahnmo and MaMoni HSSproject in the selected districts, Sylhet and Lakshmipur,respectively. Projahnmo is a partnership of the JohnsHopkins University with the Bangladesh MoHFW andBangladeshi NGOs. Projahnmo has been working in Syl-het since 2001 and has extensive experience with design-ing and evaluating newborn and maternal healthinterventions [18]. Projahnmo provided support to the im-plementation of the PSBI guidelines in two sub-districts ofSylhet: Zakigonj and Kanaighat. The USAID-fundedMaMoni HSS project is implemented in six districts ofBangladesh with the goal of improving utilization of inte-grated maternal, newborn, child health, family planning,and nutritional services [19]. The project inputs are pri-marily focused on improving the performance and cap-acity of health services at the district level. Since 2003,MaMoni HSS project has been working in all upazilas inLakshmipur to strengthen district-level health systemsand promote scale-up of maternal, neonatal and childhealth, family planning, and nutrition (MNCHFPN) inter-ventions [19]. For this study, MaMoni HSS provided sup-port to the implementation of the PSBI guidelines in onesub-district of Lakshmipur (Ramgonj).Both Projahnmo and MaMoni HSS partnered with the

MoHFW to facilitate program trainings, ensure drugavailability, and conducted joint supervision visits withthe MoHFW to the first-level facilities targeted for im-plementation. The measurement and evaluation compo-nent of this study was led by Projahnmo, with supportof MaMoni HSS in Lakshmipur, utilizing a mixed-methods approach to assess the following implementa-tion research objectives:

1. Examine feasibility of implementation of thenewly developed infection management guidelines


in young infants at UH&FWCs throughoutpatient services when referral is not accepted

2. Assess the acceptability of infection managementservices delivered on an outpatient basis atUH&FWCs among the parents and families ofyoung infants

3. Measure caregiver’s knowledge and coverage ofinfection management for young infants

4. Assess the compliance of the families to the referraladvice and new treatment regimen for young infantinfections delivered at UH&FWCs

5. Document the safety of the injectable antibiotictherapies delivered at union level facilities as pernational guidelines for infants classified as clinicalsevere infection who refuse referral advice

6. Identify barriers and facilitating factors to theimplementation of the protocol, and developstrategies to address barriers to be incorporatedinto national scale-up plans

Implementation support to the MoHFWTraining on the guidelines for outpatient management ofyoung infants with PSBIIn coordination with the MoHFW and BangabandhuSheikh Mujib Medical University (BSMMU), Projahnmoand MaMoni HSS facilitated a training-of-trainers fordistrict- and upazila-level service providers (e.g.,SACMO, FWV, FPI). Additionally, implementation sup-port teams organized orientation of both governmentand program-supported health workers and volunteerswithin the community to promote identification of dan-ger signs and the referral of cases identified in the com-munity to sub-district and union-level facilities. Theteam also supported the training of FPIs on identifica-tion of potential infection cases, referrals, and follow-upof sick young infants in the community. Refresher train-ings were provided to improve the quality of recordkeeping, PSBI case management, referral, and follow-upby SACMOs and FPIs.

Fig. 1 Map of Bangladesh highlighting implementation research study area districts


Support to monitoring and supervision of UH&FWCprovidersDistrict and sub-district level MoHFW managers wereresponsible for routine supervision and monitoring ofSACMOs and FPIs. Both Projahnmo and MaMoniHSS facilitated joint supervision visits with localMoHFW managers at UH&FWCs within the studyareas. During these visits, implementation supportteams joined the managers in their supervision ofSACMOs to observe the quality of supervision anddiscuss and resolve challenges with both supervisorsand providers in real-time. The support team pro-vided on-the-job training and mentoring to SACMOsfocusing on PSBI management, record keeping, andmonitoring. MaMoni HSS also attended the monthlymeetings for the SACMOs at the UHC in Ramgonj tosupport preparation of monthly reports.

Supply of drugs, equipment, and logistics required for PSBImanagementThe implementation-support teams coordinated withthe MoHFW and used project’s discretionary funds toprocure essential drugs, equipment, and logistics dur-ing the initial implementation period. They workedwith the MoHFW to procure the necessary drugs andsupplied them through government channels for aninterim period while the system for supplies throughthe MoHFW was being worked out.

Care-seeking message dissemination by ProjahnmoCommunity Health Workers (CHW) through home visitsAs part of other projects ongoing in Sylhet under Pro-jahnmo, there was an existing cadre of CHWs providinghome visits to mothers, newborns, and children once inevery 2 months. CHWS are local women with at leasttenth grade education, who receive 6 weeks of basichealth training, and each CHW serve a population ofabout 4000 persons. Projahnmo conducted a 1-day train-ing in the first months of implementation to orientCHWs on the updated guidelines for management of in-fections in young infants. The CHWs promoted identifi-cation of danger signs in infants and disseminated thefollowing messages: (1) when illness is identified, care-givers should take sick young infants to sub-district hos-pitals and (2) if they were unable to go to the hospital,they should seek care for the infant at the UH&FWC.

Promotion of care-seeking and referrals through ExpandedProgram on Immunization (EPI) and satellite sessionsFamily Welfare Assistants (FWA) and Health Assistants(HA) are the government frontline health workers whoconduct home visits and register pregnancies and new-borns as a part of their routine responsibilities. FWAs

and HAs received a 1-day training on the available ser-vices for treatment of PSBI for young infants at theUH&FWC. FWAs and HAs were trained to disseminatethis message to the mothers in the community duringtheir regular home visits, EPI, and satellite sessions.

Engaging community volunteers and village doctors topromote care-seeking and referralsCommunity Volunteers (CV) (1 for 250 population) ofMaMoni HSS project were oriented on newborn dangersigns, availability of sick child management services of-fered at the UH&FWC, and appropriate referral. Theydisseminated these messages within their communitiesthrough a monthly Community Action Group meeting(CAG). In addition to awareness development, theseCVs interface with community level GOB health workers(e.g., HA, FWA) at community microplanning meetingsheld monthly at the outreach EPI center. CVs supportthe MoHFW frontline health workers to gather informa-tion of births, maternal, or newborn deaths and refersick newborns in their area. Additionally, MaMoni HSSproject oriented the village doctors on identification ofPSBI cases and referred them to SACMOs as they areoften the first point of care for sick infants at the com-munity level.

Engagement of community groups for improving care-seeking and referrals in the communityProjahnmo study staff also oriented members of com-munity groups in Sylhet on newborn danger signs, theimportance of care-seeking, and the new services avail-able at the UH&FWC. Community groups are the localgoverning body for community clinics, which are thelowest tier government facility providing primary health-care on an outpatient basis to a catchment area of about6,000 in population [20]. The community group meetsperiodically to discuss the progress, challenges, and localsolutions at their forum. The Projahnmo team orientedcommunity group members to disseminate these aware-ness messages among mothers, caregivers, and othercommunity members to bolster care-seeking and com-munity referrals for sick infants.

Implementation research methodsOver the course of the 1-year implementation researchstudy (September 2015–August 2016), we conducted anevaluation, independent of the implementation support,which employed mixed-methods data collection activ-ities in 19 unions located in two sub-districts of Sylhet(9 unions) and one sub-district of Lakshmipur (10unions). A convergent parallel mixed-methods designwas used to guide quantitative and qualitative data col-lection, analysis, and interpretation of study results.Quantitative data were collected through rolling


household surveys, periodic health facility assessments,weekly extraction of data from health facility records ofyoung infants, and continuous follow-up surveys withcaregivers of infection cases in the community. Qualita-tive data were collected through process documentationactivities, in-depth interviews with senior level programimplementers, in-depth interviews (IDI), and focusgroup discussions (FGD) with UH&FWC service pro-viders, and IDI and FGD with caregivers. Both quantita-tive and qualitative data activities were used to assesseach study objective (Table 2).

Quantitative Data Collection and Sample SizeThe health facility checklist was developed in collabor-ation with study partners based on the updatedBangladesh guidelines for PSBI management, which fo-cuses on capturing health systems data on service avail-ability, general service readiness, and service-specificreadiness [15, 21]. The evaluation team piloted thechecklist in July 2015 and adapted questions prior tobaseline data collection. The baseline checklist was ad-ministered prior to the government’s rollout of theguidelines in 31 selected health facilities in Sylhet andLakshmipur. The baseline checklist assessed facilityreadiness to implement the new guidelines includingthe availability of staff, drugs, and equipment.UH&FWCs were excluded if the SACMO post was va-cant at the time of the baseline health facility assess-ment. A total of 9 UH&FWCs were selected in

Zakigonj and Kanaighat, Sylhet, and 10 UH&FWCswere selected in Ramgonj, Lakshmipur. The health fa-cility checklist was administered at two additional timepoints during the study period, 4 months after the startof implementation and then at the end of the study(August 2016). Data collectors also visited theUH&FWC weekly to abstract data from facility recordson the number of young infants that sought services.This activity provided utilization data including thenumber of young infants classified with signs of in-fection, frequency of follow-up, and treatmentreceived.Rolling household surveys were administered to ex-

plore infant illness and care-seeking history, maternalknowledge, and maternal perception of severity ofdanger signs. Household screening and the surveywere conducted by a trained group of CHWs in thestudy areas from November 2015-August 2016.CHWs recruited for this study identified all recentlydelivered women and their live born babies (0–59days) in the included catchment areas by visiting allhouseholds during the two monthly scheduled homevisits. Only married women of reproductive age(MWRA) (13–49 years) having a live birth as a preg-nancy outcome and residing in the selected unionsduring the study period were eligible to participate inthe household survey. It took approximately 2–3months to screen and administer the survey in all the19 UH&FWC catchment areas. Thus, a MWRA with

Table 2 Data collection activities by study objective

Study objective Data collection activities

Quantitative Qualitative

Examine feasibility of implementation of the newlydeveloped infection management guidelines inyoung infants at UH&FWCs through outpatientservices when referral is not accepted

• Health facility assessment • IDI & FGD with UH&FWC serviceproviders

• IDI with MoHFW programimplementers

• Process documentation ofimplementation support activities

Assess the acceptability of infection managementservices delivered on an outpatient basis at UH&FWCsamong the parents and families of young infants

• Follow-up surveys with caregivers of infectioncases in the community

• IDI with caregivers of infection cases

Measure caregiver’s knowledge and coverage ofinfection management for young infants

• Household survey with caregivers of younginfants

• FGD with caregivers of young infants

Assess the compliance of the families to the referraladvice and new treatment regimen for young infantinfections delivered at UH&FWCs

• Weekly review of young infant records at UH&FWC• Follow-up surveys with caregivers of infectioncases in the community


Document the safety of the injectable antibiotictherapies delivered at union level facilities as pernational guidelines for infants classified as clinicalsevere infection who refuse referral advice

• Weekly review of young infant records atUH&FWC

• Follow-up surveys with caregivers of infectioncases in the community


Identify barriers and facilitating factors toimplementation of the protocol, and developstrategies to address barriers to be incorporatedinto national scale-up plans

• Health facility assessment• Follow-up surveys with caregivers of infectioncases in the community

• IDI & FGD with UH&FWC serviceproviders

• IDI with MoHFW programimplementers

• Process documentation ofimplementation support activities


a young infant could only be enrolled in the surveyonce during the study period. The household surveyquestionnaire was developed utilizing questions fromprevious household surveys, which were administeredin the Sylhet study area and published by the Pro-jahnmo research group [18, 22, 23]. The populationsize in our study areas is ~ 250,000 in each of thedistricts with an annual birth cohort of 6250 (2.5%CBR). Based on our previous estimates, we expectedthat 38% of young infants will be sick as per mother’sreports for at least once in the first 2 months of theirlife [24]. Applying this estimate, there will be ~ 2375cases of reported sickness in a 1-year period in eachof the areas.We estimated current rates of care-seeking for sick

young infants from union health facilities to be 4%based on previous data. We required 76 sick infantsin each round of the survey to estimate the increaseof care-seeking from union facilities from 4% at base-line to 20% at end line with 80% power. Our primaryoutcome was reported sickness and care-seeking inthe preceding 14 days on the day of survey. Allwomen that delivered a live birth in the preceding 60days of the date of survey were asked to participatein the morbidity and care-seeking surveys if caregiverswere able to recall illness episodes. Applying 38% cu-mulative incidence of reported infant illness in thefirst 2 months of life, we expected 9% of caregivers toreport infant illness in the 14-day recall period. Thus,we targeted 845 caregivers of young infants to identify76 sick infants in each round of the survey. Assuminga response rate of 80%, which allows for an estimated20% rate of refusal or caregiver absence at the timeof the household visit, we targeted an estimated 1055caregivers of young infants per round of the survey(every 2–3 months). In order to achieve this targetsample size, we screened all women of reproductiveage for inclusion in the survey throughout the studyperiod.The study team also aimed to follow-up all young in-

fants managed under the updated guidelines to assesscompliance with follow-up, treatment outcomes, andsafety of the regimen. Facility utilization data were col-lected through weekly review of the sick infant registersat the UH&FWC by our study team to assess the num-ber of infants classified with infection, referrals, andtreatment data. The study team used these records toidentify young infants for follow-up in the community.To measure treatment compliance assuming 50% com-pliance rate with 10% precision and accounting for 10%loss to follow-up, we required complete data fromfollow-up with 107 young infants treated for infection ineach study area. Assuming 12% average care-seeking, weestimate that about 285 sick young infants will seek care

from union level facilities (12% of 2375 expected cases).However, we aimed to follow-up all young infants diag-nosed and managed under the new infection manage-ment guidelines to measure the safety of the program,which will also provide compliance data. Follow-up ofsick young infants was continuous throughout the studyperiod as the aim was to follow-up all young infant diag-nosed with infection.The total sample size requested for this study was esti-

mated at 13,590 subjects. To obtain this sample size inthe community, we obtained permission to screen 50,000 women of reproductive age in each study area total-ing 100,000 women of reproductive age during the studyperiod.

Qualitative data collection and samplingQualitative data collection took place concurrentlyduring the study period to assess program feasibilityand acceptance of the guidelines among MoHFWproviders, managers responsible for program imple-mentation, and caregivers of young infants. Amongproviders, perceptions of PSBI treatment at first-levelfacilities were collected using semi-structured IDIswith SACMOs and FPIs. We also conducted FGDswith FWVs. The SACMO at each of the selectedUH&FWCs was asked to participate in at least one,but no more than two IDIs during the study period.Interviewers asked SACMOs and FWVs about theirexperience with the guidelines, opinions on trainingand routine supervision, and facility functioning. IDIswith FPIs were conducted in the last round of datacollection (June–August 2016) to explore challengeswith follow-up of infants in the community.A subset of caregivers was selected from the list of all

caregivers with young infants identified in the study areafor FGDs as part of the qualitative component of thisstudy. We aimed to explore community perceptions ofyoung infant illness, care-seeking behaviors for illnessepisodes, and perceptions of care at the UH&FWC. ForFGDs, caregivers were selected through conveniencesampling of mothers (13–49 years) of infants under 6months of age who were willing and able to share theirexperiences with care-seeking for infant illness. Thenumber of participants for each focus group rangedfrom six to eight mothers.Caregivers of sick young infants receiving outpatient

treatment for PSBI were followed up in the communityto assess treatment compliance. We aimed to conductin-depth interviews with a subset of 30 of these care-givers in each study area throughout the study period.We purposively selected caregivers for interviews basedon their infant’s categorization of infection. We con-ducted IDIs with caregivers of infants for each categoryof infection (i.e., critical illness, clinical severe infection,


fast breathing as a single sign, and local bacterial infec-tion). The goal of these interviews was to assess thecaregiver’s experience with outpatient treatment, andreasons for non-compliance to the prescribed treatmentand follow-up visits.

Stakeholder workshopsThis implementation research study adopted anadapted action learning cycle approach, or a “plan-do-study-act” (PDSA) cycle, also known as the DemingCycle [25, 26]. According to the PDSA approach, pro-gram implementation was studied periodically, whichprovided implementers with an opportunity to identifyand address implementation challenges in real-time.With each cycle, data were collected on the programimplementation strengths and challenges and werereviewed by a group of stakeholders. The stakeholdersthen developed solutions to address the challengesidentified in the previous cycle and implemented thesechanges in the subsequent cycle (Fig. 2). The successesand challenges of the revised program approach werestudied in the subsequent cycle. We arranged a stake-holders’ meeting following each round of data collec-tion, during which the preliminary results werereviewed, and stakeholders assessed the implementationprogress and challenges. The records from the stake-holder meetings served as documentation of the pro-gram learning and were reported alongside the resultsfrom data collection activities. The evaluation teamworked closely with the implementation support team

and the MoHFW implementers to perform all the pre-paratory works and organized stakeholder review meet-ings for sharing and gathering inputs.A total of two stakeholder workshops were held in

Dhaka after the first and third round of data collec-tion, in January and September 2016, respectively.These workshops aimed to bring together implemen-tation and study partners, district level officials, andstakeholders at the central level to share findingsrelated to both successes and challenges that aroseduring the process of implementing the new guide-lines. Through these workshops, participants sharedearly learnings from implementation support andevaluation activities and worked together to developsolutions to better support the implementation of thenew guidelines.

Data AnalysisQuantitative data were entered and stored in Micro-soft SQL server and analyzed using Stata Special Edi-tion 14 (College station, Texas, USA) [27]. Theanalysis plan for this data included summary statisticsof distribution and cross-tabulation of indicators usingthe appropriate tests for significance (e.g., Student’s ttest and chi-square). Qualitative data were analyzedfollowing an adapted Framework approach [28] foridentification of inductive and deductive themes. Acodebook was developed to ensure consistency in thebroader thematic concepts that was sought in thedata. Johns Hopkins University (JHU) qualitative

Fig. 2 Adapted “plan-do-study-act” cycle including study activities at each stage. This implementation research study adopted an adapted actionlearning cycle approach, or a “plan-do-study-act” (PDSA) cycle [25, 26] to guide program learning and inform adjustments toimplementation support


researchers applied thematic codes systematically tothe data and examined for patterns. The interviewswere transcribed and translated into English. TheJHU qualitative team coded the English transcriptsand analyzed as per a framework for analysis basedon the objectives of the program and itsimplementation.

Ethical approvalWe obtained ethical clearance to conduct the study fromethical review committee and/or internal review boardsof Bangladesh Institute of Child Health and Johns Hop-kins Bloomberg School of Public Health. They reviewedand approved the research plan, consent forms, and datacollection forms.

DiscussionThis article describes the design of an implementationresearch study, which included support to theBangladesh MoHFW to implement revised guidelinesfor the management of young infants suffering fromPSBI and a mixed-methods evaluation. The evaluationaimed to identify facilitators and barriers to the imple-mentation of the guidelines in first-level health facilitiesto inform scale-up. The WHO guidelines are intendedto be adopted by national governments and imple-mented by health workers in limited resource settings.Thus, there is a need to study how these guidelines willbe implemented outside of randomized controlled trials.Implementation research provides an opportunity tounderstand what, why, and how interventions work inreal-world conditions [29]. Our incorporation of imple-mentation research outcomes provides the opportunityfor us to assess why the program was successful or un-successful in meeting goals, which will be valuablefeedback for both the MoHFW, WHO, USAID, andother global stakeholders [30–32]. WHO is coordinatingadditional implementation research studies for PSBIguideline rollout in Pakistan, India, Nigeria, Malawi,DRC, and Ethiopia. Findings from this study will be dis-seminated among program managers, policy-makers, de-velopment partners, and other stakeholders.The strength of this study is the use of both quantita-

tive and qualitative approaches to provide a deeper un-derstanding of the research questions than eithermethod separately [33, 34]. This approach is well-suitedto the implementation research because it provides away to understand multiple perspectives and multipleoutcomes grounded within local context [29, 35].Given the lack of a control group and short study

period, it will not be possible to causally link implemen-tation support activities to observed changes in thepopulation. For the household survey, this limitation isexacerbated because we do not have survey data

collected prior to the MoHFW’s rollout of the PSBIguidelines in the study areas. It is important to note thatthe lack of a comparison group and lack ofrandomization make the study more vulnerable to in-ternal and external threats to validity. We aimed to im-prove internal validity by collecting data at multiplepoints in time. However, our study period was limited to1-year, which was necessary based on the GOB’s plansfor scale-up.The potential lack of generalizability of these study

findings to other developing country settings is anotherlimitation of this study. Although generalizability wasnot a primary goal for this study, it will be important toconsider this when formulating conclusions. This imple-mentation research study focuses on implementation re-search outcomes in the Bangladesh health system, thusfindings will not be directly transferable to other coun-tries. Given that this study is being conducted prior tonational scale-up of the program, it will also be import-ant to consider the generalizability of the findings toother areas in the country. Both Sylhet and Lakshmipurhave well-established, large-scale programs aimed to im-prove maternal, newborn and child health. As a result, itwill be difficult to tease out the improvements in mater-nal knowledge or care-seeking that may be linked to thecommunity mobilization activities. When formulatingstudy conclusions, it will be important to describe theother programs operating in each area and the impactthese programs may have on study findings.

AbbreviationsBSMMU: Bangabandhu Sheikh Mujib Medical University; CAG: CommunityAction Group; CBR: Crude birth rate; CHW: Community health workers;CNCP: Comprehensive Newborn Care Package; CV: Community volunteer;DRC: Democratic Republic of Congo; EPI: Expanded Program onImmunization; FGD: Focus group discussion; FPI: Family Planning Inspector;FWA: Family Welfare Assistant; FWV: Family Welfare Visitor; GOB: Governmentof Bangladesh; HA: Health Assistant; IDI: In-depth interviews; IMCI: IntegratedManagement of Childhood Illness; JHU: Johns Hopkins University; MaMoniHSS: MaMoni Health Systems Strengthening; MNCHFPN: Maternal, Neonataland Child Health, Family Planning and Nutrition; MoHFW: Ministry of Healthand Family Welfare; MWRA: Married women of reproductive age; PDSA: Plan-do-study-act cycle; PSBI: Possible serious bacterial infection; SACMO: Sub-Assistant Community Medical Officer; SC/SNL: Save the Children/SavingNewborn Lives; UH&FWC: Union Health and Family Welfare Center;UHC: Upazila Health Complex; USAID: United States Agency for InternationalDevelopment; WHO: World Health Organization

AcknowledgementsThe authors thank the study participants and Projahnmo field staff for theirefforts in implementing the study, and the Ministry of Health and FamilyWelfare, Government of Bangladesh, for leading program implementationand for their support and collaboration in all phases of the study. We alsothank our study partners with USAID’s MaMoni Health System Strengthening(HSS) project, implemented by Save the Children Bangladesh; Johns HopkinsBloomberg School of Public Health; and icddr,b for their valuable inputs.

Authors’ contributionsAB conceptualized and designed this implementation research study. Allauthors were involved in project implementation and/or design of studyprocedures. SA, MM, NB, SA, and JG undertook the project in Bangladeshand led the field teams in supporting program implementation and data


collection. JA and AB wrote the first draft of the paper. All authors reviewedand approved the final version of the manuscript

FundingThis study was supported by United States Agency for InternationalDevelopment (USAID) through the Health Research Challenge for Impact(HRCI) Cooperative Agreement (#GHS-A-00-09-00004-00). The contents arethe responsibility of the authors and do not necessarily reflect the views ofUSAID or the United States Government.

Availability of data and materialsNot applicable

Ethics approval and consent to participateWe obtained ethical clearance to conduct the study from ethical reviewcommittee and/or internal review boards of Bangladesh Institute ofChild Health and Johns Hopkins Bloomberg School of Public Health.They reviewed and approved the research plan, consent forms and datacollection forms.

Consent for publicationNot applicable

Competing interestsThe authors declare that they have no competing interests.

Author details1Johns Hopkins University-Bangladesh, Dhaka 1213, Bangladesh.2International Center for Maternal and Newborn Health, Department ofInternational Health, Johns Hopkins Bloomberg School of Public Health,Baltimore, MD 21205, USA. 3Department of Public Health, School of Healthand Life Sciences, North South University, Dhaka 1229, Bangladesh.4Department of Sociology, Anthropology, and Health Administration andPolicy, University of Maryland, Baltimore County, Baltimore, MD, USA.5Jhpiego, Baltimore, MD, USA. 6BRAC James P Grant School of Public Health,BRAC University, Dhaka, Bangladesh. 7USAID’s MaMoni Health SystemsStrengthening Project, Save the Children, Washington, DC, USA. 8Maternaland Child Health Division, icddr,b, Dhaka 1212, Bangladesh.

Received: 13 May 2018 Accepted: 29 October 2019

References1. United Nations Children’s Fund. Levels & trends in child mortality: report

2015. In: Estimates developed by the UN Inter-agency Group for ChildMorality Estimation. New York: United Nations Children’s Fund; 2015.

2. Liu L, et al. Global, regional, and national causes of under-5 mortality in2000-15: an updated systematic analysis with implications for theSustainable Development Goals. Lancet. 2016;388(10063):3027–35.

3. Lawn JE, Cousens S, Zupan J. 4 million neonatal deaths: when? where?why? Lancet. 2005;365(9462):891–900.

4. Liu L, et al. Global, regional, and national causes of child mortality in 2000-13, with projections to inform post-2015 priorities: an updated systematicanalysis. Lancet. 2015;385(9966):430–40.

5. Baqui AH, et al. Effect of community-based newborn care on cause-specificneonatal mortality in Sylhet district, Bangladesh: findings of a cluster-randomized controlled trial. J Perinatol. 2016;36(1):71–6.

6. BDHS, Bangladesh Demographic and Health Survey 2011. 2013, NationalInstitute of Population Research and Training (NIPORT), Mitra andAssociates, and Macro International: Dhaka, Bangladesh and Calverton,Maryland, USA.

7. Al-Eidan FA, et al. Sequential antimicrobial therapy: treatment of severelower respiratory tract infections in children. J Antimicrob Chemother. 1999;44(5):709–15.

8. Peterson S, et al. Coping with paediatric referral--Ugandan parents’experience. Lancet. 2004;363(9425):1955–6.

9. Qazi SA, et al. An innovative multipartner research program to addressdetection, assessment and treatment of neonatal infections in low-resourcesettings. Pediatr Infect Dis J. 2013;32:S3–6.

10. Zaidi AK, et al. Scientific rationale for study design of community-basedsimplified antibiotic therapy trials in newborns and young infants with

clinically diagnosed severe infections or fast breathing in South Asia andsub-Saharan Africa. Pediatr Infect Dis J. 2013;32:S7–S11.

11. Baqui AH, et al. Safety and efficacy of alternative antibiotic regimenscompared with 7 day injectable procaine benzylpenicillin andgentamicin for outpatient treatment of neonates and young infantswith clinical signs of severe infection when referral is not possible: arandomised, open-label, equivalence trial. Lancet Global Health. 2015;3(5):e279–87.

12. Mir F, et al. Simplified antibiotic regimens for treatment of clinicalsevere infection in the outpatient setting when referral is not possiblefor young infants in Pakistan (Simplified Antibiotic Therapy Trial [SATT]):a randomised, open-label, equivalence trial. Lancet Global Health. 2017;5(2):e177–85.

13. Tshefu A, et al. Simplified antibiotic regimens compared with injectableprocaine benzylpenicillin plus gentamicin for treatment of neonates andyoung infants with clinical signs of possible serious bacterial infection whenreferral is not possible: a randomised, open-label, equivalence trial. Lancet.2015;385(9979):1767–76.

14. World Health Organization, Guideline: Managing possible serious bacterialinfection in young infants when referral is not feasible. Geneva: WorldHealth Organization; 2015.

15. National Technical Working Committee on Newborn Health, BangladeshNational Guideline: Management of infection of the 0-59 days infants atunion level facilities and NGO clinics without indoor facilities. 2015.

16. Seddiky MA, Rahman ST. Role of Union Health and Family Welfare Center(UH&FWC) to promote maternal education and reduce child mortality ratein Bangladesh. J Emerg Trends Educ Res Policy Stud. 2015;6(2):169–81.

17. Ministry of Health and Family Welfare Bangladesh. Health Bulletin 2016.2016 [Accessed 2 Feb 2017]; Available from: http://www.dghs.gov.bd/images/docs/Publicaations/HB%202016%20_2nd_edition_13_01_17.pdf.

18. Baqui AH, et al. Effect of community-based newborn-care interventionpackage implemented through two service-delivery strategies in Sylhetdistrict, Bangladesh: a cluster-randomised controlled trial. Lancet. 2008;371(9628):1936–44.

19. MCHIP. MaMoni Health Systems Strenghtening Project-Bangladesh.[Accessed 6 Apr 2017]; Available from: http://www.mchip.net/node/89.

20. World Health Organization. Community clinics in Bangladesh: Bringinghealth care to the doorstep of rural people. 2017 [Accessed 2 Feb 2017];Available from: http://www.searo.who.int/mediacentre/events/community-clinic-bangladesh-story.pdf?ua=1.

21. World Health Organization, Service availability and readiness assessment(SARA), in Implementation Guide, version 2.2. Geneva: World HealthOrganization; 2015.

22. Baqui AH, et al. Levels, timing, and etiology of stillbirths in Sylhet district ofBangladesh. BMC Pregnancy Childbirth. 2011;11:25.

23. Baqui AH, et al. Rates, timing and causes of neonatal deaths in rural India:implications for neonatal health programmes. Bull World Health Organ.2006;84(9):706–13.

24. Mitra DK, et al. Implementation of the ANISA protocol in Sylhet, Bangladesh:Challenges and Solutions. Pediatr Infect Dis J. 2016;35(5 Suppl 1):S55–9.

25. Brassard M, Ritter D. The memory Jogger II: A Pocketguide of Tools forContinuous Improvement & Effective Planning. Goal/QPC; 1988.

26. Peters HD, Tran TN, Adam T. Implementation research in health: a practicalguide. Geneva: Alliance for Health Policy and Systems Research, WorldHealth Organization; 2013.

27. StataCorp, Stata Statistical Software: Release 14. 2015, StataCorp LP: CollegeStation, TX.

28. Ritchie J, Spencer L. In: Bryman A, Burgess R, editors. Qualitative dataanalysis for applied policy research, in Qualitative data analysis for appliedpolicy research. London: Routledge; 1994. p. 173–94.

29. Peters DH, et al. Implementation research: what it is and how to do it. BMJ.2013;347:f6753.

30. Habicht JP, Victora CG, Vaughan JP. Evaluation designs for adequacy,plausibility and probability of public health programme performance andimpact. Int J Epidemiol. 1999;28(1):10–8.

31. Peters DH. The application of systems thinking in health: why use systemsthinking? Health Res Policy Syst. 2014;12:51.

32. Victora CG, Habicht JP, Bryce J. Evidence-based public health: movingbeyond randomized trials. Am J Public Health. 2004;94(3):400–5.

33. Creswell J, Clark VP. Designing and Conducting Mixed Methods Research.2nd Edition ed. 2011. USA: SAGE Publication Inc.


http://www.dghs.gov.bd/images/docs/Publicaations/HB%202016%20_2nd_edition_13_01_17.pdfhttp://www.dghs.gov.bd/images/docs/Publicaations/HB%202016%20_2nd_edition_13_01_17.pdfhttp://www.mchip.net/node/89http://www.searo.who.int/mediacentre/events/community-clinic-bangladesh-story.pdf?ua=1http://www.searo.who.int/mediacentre/events/community-clinic-bangladesh-story.pdf?ua=1

34. Teddlie C. Tashakkori A. Foundations of mixed methods research:Integrating quantitative and qualitative approaches in the social andbehavioral sciences. London: Sage; 2009.

35. Proctor E, et al. Outcomes for implementation research: conceptualdistinctions, measurement challenges, and research agenda. Adm PolicyMent Health. 2011;38(2):65–76.

Publisher’s NoteSpringer Nature remains neutral with regard to jurisdictional claims inpublished maps and institutional affiliations.


AbstractBackgroundImplementation supportMethodsDiscussion

BackgroundMethodsStudy settingStudy designImplementation support to the MoHFWTraining on the guidelines for outpatient management of young infants with PSBISupport to monitoring and supervision of UH&FWC providersSupply of drugs, equipment, and logistics required for PSBI managementCare-seeking message dissemination by Projahnmo Community Health Workers (CHW) through home visitsPromotion of care-seeking and referrals through Expanded Program on Immunization (EPI) and satellite sessionsEngaging community volunteers and village doctors to promote care-seeking and referralsEngagement of community groups for improving care-seeking and referrals in the community

Implementation research methodsQuantitative Data Collection and Sample SizeQualitative data collection and samplingStakeholder workshopsData AnalysisEthical approval

DiscussionAbbreviationsAcknowledgementsAuthors’ contributionsFundingAvailability of data and materialsEthics approval and consent to participateConsent for publicationCompeting interestsAuthor detailsReferencesPublisher’s Note

Implementation research to support Bangladesh Ministry of ... · F or 35.5 ° C † Raised temperature (> 99.5 ° F or 37.5 ° C) † Less movement/movement only when stimulated †

Documents