The Author(s) BMC Infectious Diseases 2017, 17(Suppl 1):707DOI 10.1186/s12879-017-2770-5

RESEARCH Open Access

Optimising diagnosis of viraemic hepatitisC infection: the development of a targetproduct profile

Elena Ivanova Reipold1*, Philippa Easterbrook4, Alessandra Trianni1, Nivedha Panneer2, Douglas Krakower3,Stefano Ongarello1, Teri Roberts1, Veronica Miller2 and Claudia Denkinger1,3

Abstract

Background: The current low access to virological testing to confirm chronic viraemic HCV infection in low- andmiddle-income countries (LMIC) is limiting the rollout of hepatitis C (HCV) care. Existing tests are complex, costlyand require sophisticated laboratory infrastructure. Diagnostic manufacturers need guidance on the optimalcharacteristics a virological test needs to have to ensure the greatest impact on HCV diagnosis and treatment inLMIC. Our objective was to develop a target product profile (TPP) for diagnosis of HCV viraemia using a globalstakeholder consensus-based approach.

Methods: Based on the standardised process established to develop consensus-based TPPs, we followed five keysteps. (i) Identifying key potential global stakeholders for consultation and input into the TPP development process.(ii) Informal priority-setting exercise with key experts to identify the needs that should be the highest priority forthe TPP development; (iii) Defining the key TPP domains (scope, performance and operational characteristics andprice). (iv) Delphi-like process with larger group of key stakeholder to facilitate feedback on the key TPP criteria andconsensus building based on pre-defined consensus criteria. (v) A final consensus-gathering meeting for discussionsaround disputed criteria. A complementary values and preferences survey helped to assess trade-offs betweendifferent key characteristics.

Results: The following key attributes for the TPP for a test to confirm HCV viraemic infection were identified: Thescope defined is for both HCV detection as well as confirmation of cure. The timeline of development for testsenvisioned in the TPP is 5 years. The test should be developed for use by health-care workers or laboratorytechnicians with limited training in countries with a medium to high prevalence of HCV (1.5–3.5% and >3.5%) andin high-risk populations in low prevalence settings (<1.5%). A clinical sensitivity at a minimum of 90% is consideredsufficient (analytical sensitivity of the equivalent of 3000 IU/ml), particularly if the test increases access to testingthrough an affordable price, increase ease-of-use and feasibility on capillary blood. Polyvalency would be optimal(i.e. ability to test for HIV and others). The only characteristic that full agreement could not be achieved on was theprice for a virological test. Discussants felt that to reach the optimal target price substantial trade-offs had to bemade (e.g. in regards to sensitivity and integration).

Conclusion: The TPP and V&P survey results define the need for an easy-to-use, low cost test to increase access todiagnosis and linkage to care in LMIC.

* Correspondence: elena.ivanova@finddx.org1FIND, MSF Access Campaign, Geneva, SwitzerlandFull list of author information is available at the end of the article

© The Author(s). 2017 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.

The Author(s) BMC Infectious Diseases 2017, 17(Suppl 1):707 Page 160 of 175

BackgroundHepatitis C virus (HCV) causes both acute and chronic in-fection, and is a leading cause of liver-related mortalityworldwide [1]. Spontaneous clearance of acute HCV infec-tion occurs in between 15 and 45% of infected individualsin the absence of treatment [2]. The remaining 55–85% ofpersons who do not clear HCV within six months are de-fined as having chronic viraemic infection and requiretreatment to prevent the substantial liver-related morbid-ity and mortality. It is estimated that 71 million people areliving with chronic viraemic HCV infection [3].The treatment landscape for HCV has undergone a dra-

matic transformation in recent years, from the use of com-plex interferon-based regimens with low cure rates andsignificant rates of adverse effects, to availability of shortall-oral direct acting antiviral (DAA) regimens, which areboth well tolerated and offer very high cure rates of >90%across different HCV genotypes after 12 weeks of treatment[4]. There has also been recent expansion of access in low-and middle income countries (LMIC) as a result of theavailability of generic DAAs [5–7] resulting in substantialprice reductions and price of US $ 100–250 for a 2-drugtreatment course is within reach [8, 9].Despite these opportunities for treatment, a major

bottleneck to appropriate HCV care is diagnosis [10]. It isestimated that around 20% of infected persons globallyhave been tested and are aware of their diagnosis [3], butthis is much lower in LMICs [11]. There are several keyreasons for this low uptake: lack of awareness about HCVamong population, exiting diagnostic algorithms are com-plex, laboratory capacity to diagnose hepatitis C in LMICremains very low and is largely located in the private sec-tor, and available tests are costly [12–14]. The currentdiagnostic algorithm first requires screening for evidenceof past or current HCV infection with a serological assaythat detects hepatitis C antibody (either an enzyme im-munoassay (EIA) or by a rapid diagnostic test (RDT)).Those with positive serology will need further testing toconfirm the presence of HCV viraemia, and need for treat-ment, by nucleic acid testing (NAT) detection HCV RNA(either quantitative or qualitative) or antigen detection tar-geting HCV (p22) core antigen (HCV cAg) [15–17].Currently available commercial quantitative and qualita-

tive HCV RNA or HCV cAg assays are focused on central-ized laboratory infrastructure (e.g. Roche COBAS TaqmanHCV assay, Abbott RealTime HCV assay, HologicAptima® HCV Quant Dx Assay) [13]. While such plat-forms may play an important role in countries that favoura centralized infrastructure, and have the appropriate spe-cimen referral system in place, improvements are neces-sary on the specimen collection and transport (e.g.through Dried Blood Spots) and on reaching affordability.Recently, one integrated platform for decentralized usehas entered the market (i.e. Cepheid GeneXpert) and has

also received WHO pre-qualification [18]. Other promis-ing diagnostics are under development, but manufacturersneed further guidance on the optimal characteristics ofthe tests that can be implemented in LMIC. An HCVRNA or HCV cAg test that can be performed near or atthe site of patient care (point-of-care [POC] testing) mayhelp increase access to HCV care especially in resource-limited settings and hard to reach populations. Also, theadvantages of rapid turn-around times offered by near-patient or POC platforms may reduce loss to follow-upand allow for immediate treatment decisions [19]. Inaddition, if these tests can be made affordable, it may bepossible to dispense with the need for an initial serologicalscreening test and have a single test strategy to diagnoseviraemic infection (Fig. 1).A target product profile (TPP) is a tool that has been

used widely to support and guide drug, and diagnostic de-velopment in various fields of health [20, 21]. This articledescribes the development of a TPP for a near-patient testfor diagnosis of HCV viraemia using a global stakeholderconsensus-based approach. The TPP process was comple-mented by a comprehensive survey of values and prefer-ences (V&P survey) to seek additional perspectives on theTPP criteria and capture acceptable trade-offs betweentest price, performance and operational characteristics.This TPP is intended to inform test developers as well asdonors and provide a tool for communication in accelerat-ing the path towards HCV diagnostic tests suitable forscale-up of HCV care in LMIC.

MethodsDeveloping target product profiles for drugs anddiagnosticsA target product profile (TPP) is a strategic planning tool– usually in the form of a document that provides a clearstatement of desired outcome of product development [22].There have been several TPPs for diagnosis of tuberculosis,sexually transmitted diseases and others [21, 23–25]. ATPP aims to (i) Define the medical/public health need; (ii)Make the need transparent to test developers to increaseinterest in the field by defining features and characteristicsof the desired test; and (iii) Outline features that wouldprovide a competitive advantage, utilizing the input of keystakeholders in the field. The standard categories of criteriaincluded in a TPP are scope, performance and operationalcharacteristics and price.

Methodology for development of TPPs for HCV RNA andcore antigen testsThere were five key steps in the process of developingthis TPP.

1. Identifying key potential global stakeholders forconsultation and input into the TPP development

Fig. 1 One-step and two-step testing strategy for HCV diagnosis. RDT states for rapid diagnostic test; POC – point-of-care (i.e. test can beperformed at any level of patient care)

The Author(s) BMC Infectious Diseases 2017, 17(Suppl 1):707 Page 161 of 175

process. A list of 50 potential stakeholders wereidentified for consultation and input into the TPPdevelopment process. The stakeholders were chosento provide a representation from (i) NationalHepatitis programs; (ii) Implementers and Clinicians(including from NGOs); (iii) Technical agencies andresearchers; (iv) International organizations; (v)Advocacy organizations/Civil society; (vi) industry(pharmaceutical and IVD-manufacturers)

2. Informal priority-setting exercise with key experts.Second, we undertook an informal priority-settingexercise with key stakeholders to identify the needsthat should be the highest priority in the TPP devel-opment. Given the availability of centralized plat-forms to diagnose viraemic infection, it was decidedto focus this TPP development on point-of-care ornear-patient diagnostic tests that detect the presenceof HCV RNA or HCV cAg. Such diagnostic testswere considered to play an important role in settingswith challenges in specimen and results transfer andhigh loss-to-follow-up after initial diagnosis [26, 27].

3. Defining the key TPP domains (scope, performanceand operational characteristics and price). Third, wedefined the domains and characteristics of acomprehensive TPP. The domains included (i) Scope(characteristics: goal of the test, target population,level of the health-care system where the test couldbe run and intended end user(s) and lowest settingfor implementation); (ii) Performance (characteris-tics: clinical and analytical sensitivity and specificity,polyvalency and quantitation); (iii) Operational(characteristics: specimen type and preparation

steps, time to result, specimen through-put and cap-acity, biosafety and waste disposal, instrumentation,power requirements, maintenance calibration, dataanalysis, connectivity, result capture, operatingtemperature/humidity, reagent kit transport/storageand stability, internal quality control); (iv) price(characteristics: maximum price per test and for in-strumentation), as well as the most important per-formance and operational characteristics and prices(Table 1). These characteristics were described interms of both minimal and optimal characteristics,which defined a range, where “minimal” referred tothe lowest acceptable output for a characteristic and“optimal” to the ideal target (preferable/ideal). Theoriginal TPPs were developed by FIND and theForum for Collaborative HIV Research (“theForum”), in collaboration with WHO’s Global Hepa-titis Programme. Additional input was sought fromexperts in the field at the U.S. Center for DiseaseControl and Prevention (CDC), Médecins SansFrontières (MSF), the Treatment Action Group(TAG), and WHO’s Prequalification Program.

4. Delphi-like process with larger group of key stakeholderto facilitate feedback on the key TPP criteria andconsensus building based on pre-defined consensus cri-teria. Fourth, a Delphi process [28] was used tofacilitate anonymous feedback on the key TPP criteriaand consensus building. Organizations/individuals thatwere invited and agreed to participate were asked toprovide a statement on their level of agreement witheach of the proposed characteristics for each TPP.Agreement was scored on a Likert scale ranging from 1

Table

1Com

bine

dHCVRN

AandcA

gTPPforan

HCVtest

Characteristic

Optim

alMinim

alRatio

naleandeviden

ce

Scop

e

Goalo

ftest

Thego

alof

thetestistw

o-fold:

1.To

diagno

seactiveviraem

icHCVinfection(new

orreinfection)

andprovidebaselinevirological

assessmen

t(quantitativeor

qualitative);

2.To

confirm

cure

upon

treatm

entcompletion.

Ideally,the

testwou

ldbe

done

with

thepu

rposeof

initiatingtreatm

entwith

inthesameclinicalencoun

ter

orthesameday.

Not

intend

edforbloo

dscreen

ing.

Thetim

elineof

developm

entfortestsen

vision

edin

theTPPis5years.

Thego

alof

thetestistw

o-fold:

1.To

diagno

seactiveHCVviraem

icinfection(new

orreinfection)

andprovidebaselinevirological

assessmen

t(qualitative)

with

thepu

rposeof

initiating

treatm

ent;

2.To

confirm

cure

upon

treatm

entcompletion.

Not

intend

edforbloo

dscreen

ing.

Thetim

elineof

developm

entfortestsen

vision

edin

theTPPis5years.

Detectio

ncanbe

perfo

rmed

byHCVRN

Atestor

byHCVcA

gde

tection.Presen

ceof

HCVRN

Aor

cAgin

apatient

isindicativeof

activeHCVinfection.Currently,

theHCVRN

Aor

cAgtestispe

rform

edafterapo

sitive

anti-HCVserologicaltest(i.e.tw

o-step

algo

rithm

).Con

ceivably,p

rovide

dtheprevalen

ceissubstantial

andthecostof

theHCVRN

Aor

cAgtestislow,eith

ertestcouldbe

used

inaon

e-step

algo

rithm

.

Target

popu

latio

nCou

ntrieswith

amed

ium

tohigh

prevalen

ceof

HCV(1.5–3.5%

and>3.5%

)High-riskpo

pulatio

nsin

low

prevalen

cesettings

(<1.5%

).High-riskpo

pulatio

nsinclud

e:pe

rson

swho

inject

drug

sor

have

used

intranasaldrug

s(PWID),pe

ople

livingwith

HIV

(PLW

H),men

who

have

sexwith

men

(MSM

),prison

ers,pe

oplewith

tattoo

s,sexworkers,

peop

lewith

frequ

entcontactwith

thehe

alth-caresys-

tem

(i.e.chronically

ill)andchildrenbo

rnto

HCV-

infected

mothe

rs.Inorde

rto

achievethelong

-term

goalof

HCVelim

ination,

optim

allythetestshou

ldbe

perfo

rmed

onallp

atientsin

prim

arycare

settings,

antenatalclinicsandin

commun

ityscreen

ing

prog

rammes.

Target

operator

oftest

Com

mun

ityworkerswith

minim

altraining

Health

-careworkersor

labo

ratory

technicianswith

limitedtraining

(i.e.ableto

operatean

integrated

test

with

minim

aladditio

nalsteps)

Lowestlevelo

fsettingfor

implem

entatio

n(pub

lic&private)

Com

mun

itycentres

Districtho

spital(LevelII)

Perfo

rmance

characteristics

Diagn

ostic

sensitivity*

>99%

90%–95%

Rationaleof

optim

altest:Ideallyatestshouldbe

assensitive

andspecificas

availableplasma-basedHCV

RNA

tests.Acommon

lyused

referencestandard

istheVER-

SANTHCV

RNAQualitativeAssay,w

hich

isFD

A-approved

ford

iagn

osisof

activeHCV

infection(although

theVER-

SANTHCV

RNAQualitativeAssay

isbeingtakenoffthe

market,itremains

themostanalyticallysensitive

assayand

was

used

asthego

ldstandard

inmostinstances).

Rationaleof

minimal:Ifa

testiseasiertoimplem

entat

lowerlevelsof

thehealth

caresystem

withoutrequiring

substantialtechn

icalexpertise

orcomplex

labo

ratory

infrastructure,andislesscostly,thenacomprom

isecanbe

madeon

sensitivity.A

testwith

asubo

ptimalsensitivityof

90–95%

with

improved

operationalcharacteristicswas

considered

acceptableby

stakeholdersas

itwould

improveratesof

diagno

sissubstantially

overwhatis

currentlypo

ssible.H

owever,nostud

iesor

mod

ellinghave

been

done

ontheminimalacceptablesensitivityandthe

The Author(s) BMC Infectious Diseases 2017, 17(Suppl 1):707 Page 162 of 175

Table

1Com

bine

dHCVRN

AandcA

gTPPforan

HCVtest(Con

tinued)

Characteristic

Optim

alMinim

alRatio

naleandeviden

ce

optim

alothercharacteristicsneeded

byatestforH

CVdiagno

sisto

lead

tosubstantialimprovem

entinHCV

detectionon

apo

pulationlevel.

Mod

ellingworkforTBhasprovided

insig

htsthatcould

potentially

beapplicableforH

CVas

well.Amod

elshow

edthatforthe

WHOSoutheastAsia

Region

aPO

Cbiom

arkertestwith

asensitivityof

50%forsmear-

negativeTB,ifemployed

atthemostp

eripheralhealth

-care

setting,wou

ldresultinasim

ilarreductio

ninTB

inci-

denceas

atestwith

70%sensitivityforsmear-negative

TBthatwou

ldbe

employed

atthedistrictlevel(e.g.

XpertMTB/RIF)[17].How

ever,the

exacttrade-offb

e-tweenalower

sensitivity(fo

rsmear-negativeTB)and

anincrease

inaccessto

testingissettingde

pend

ent.

Und

ertheminim

alscen

ario,som

epatientswou

ldbe

incorrectly

diagno

sedas

nothaving

activeHCV

infection,Whatim

pact

that

wou

ldhave

onpatient

andprovider

behaviou

risun

clear.

Analyticalsensitivity

(com

parison

with

HCV

RNAtestreferencestandard)

200IU/m

l1000–3000IU/m

lAmon

gthemajority

ofinfected

individu

alswith

chronic

HCV

infection,HCV

RNAviralloads

arebetween10

4and

107IU/m

l[13].Instud

iesof

viraldynam

icsdu

ringacute

infection,viralloads

aslowas

3logIU/m

l(or

1000

IU/

ml)wereseen

duringthefirstfour

mon

thsafter

infection[13].The

optim

alLO

Dof

200IU/m

lsho

uld

thereforedetectmostpatients(>99%).Ataminimum

,analyticalsensitivityof

1000–3000IU/m

lor3

fmolcAg/l

(current

LODof

theAbb

ottHCV

cAgassay),the

corre

spon

ding

clinicalsensitivityshou

ldbe

90–95%

.Upp

erlim

itof

thedynamicrang

eshou

ldbe

equivalent

tothat

ofcurren

tlabo

ratory-based

quanti-

tativeHCVRN

Atests.AnHCVRN

Atestshou

ldbe

stan-

dardized

with

theWHOInternationalStand

ardfor

Hep

atitisCVirusRN

A,ashasbe

endo

newith

curren

tFD

A-app

rovedandCE-markedqu

alitativeandqu

anti-

tativeHCVRN

Aassays.

Interestingly,HCV

cAgtesthave

been

show

nto

benegativeam

ongapo

rtion

ofun

treated

individu

alswith

high

HCV

RNAlevels,indicatingthelikelypresence

ofmutantvariants[18].The

limitationof

thecAgassayto

accuratelydetectthesevariantsmay

presenta

challeng

eto

elimination.

Datafro

mpatientswho

relapsed

aftertreatm

entwith

peg-interfe

ronandrib

avirintherapyindicate

that

HCV

rebo

unds

quicklyto

high

levels(103

IU/m

land

greater)

with

inafew

weeks

aftertheen

dof

treatm

ent[19].

Early

data

from

DAA-based

therapysugg

eststhat

aneven

morerapidrelapsewou

ldhapp

en(unp

ublishe

ddata;com

mun

icationwith

A.H

ill).Given

thehigh

cor-

relatio

nbe

tweenHCVcA

gandHCVRN

Alevels,eith

ertestwou

ldlikelybe

suitableformon

itorin

gvirologic

The Author(s) BMC Infectious Diseases 2017, 17(Suppl 1):707 Page 163 of 175

Table

1Com

bine

dHCVRN

AandcA

gTPPforan

HCVtest(Con

tinued)

Characteristic

Optim

alMinim

alRatio

naleandeviden

ce

respon

seaftertreatm

entcompletionseveralw

eeks

aftercompletionof

therapy.

Diagn

ostic

specificity

(com

parison

with

HCV

RNAreferencestandard)*

>99%

>98%

Sincethetestisatestford

etectionof

activeHCV

infection,itshou

ldbe

asspecificas

currentcommercially

availableandFD

A-app

rovedHCV

RNAteststo

avoid

false

positiveresults.

Analyticalspecificity

–HCVde

tection

Nocrossreactivity

with

endo

geno

ussubstanceand

exog

enou

sfactors(e.g.H

IV-1,H

IV-2,H

BV,H

EV,antim

a-larials,anti-TB,ART)

Nocrossreactivity

with

endo

geno

ussubstanceand

exog

enou

sfactors(e.g.H

IV-1,H

IV-2,H

BV,H

EV,antim

a-larials,anti-TB,ART)

Polyvalency

Abilityto

detect

HIV,hep

atitisBon

thesame

instrumen

t

Quantitatio

nQuantitative

Qualitative

Treatm

entmon

itoringisno

tconsidered

necessaryor

feasiblewith

novelD

AAagents[14],therefore

aqu

alitativetestresultispreferred.According

tostakeholdero

pinion

,aqu

antitativeresultwou

ldbe

beneficialasitallowsresearch

questio

nsto

beinvestigated;how

ever,itcann

otcomeatan

increased

cost.

Ope

ratio

nalC

haracteristics

Specim

entype

Capillarywho

lebloo

dVeno

uswho

lebloo

dor

plasma

Theem

phasisisforthe

useof

capillarywho

lebloo

dthat

candiagno

seinfectionintheclinicwith

outrequ

iring

additio

nallaboratoryequipm

entsuch

asabenchtop

centrifug

e.Theneed

forp

hlebotom

yto

draw

veno

uswho

lebloo

dwou

ldlim

ittheapplicabilityof

thetestinlower

settings

ofcare

asperstakeho

lder

opinion.Ifplasmaisto

bea

specimen

type

(minimalcriteria),theplasmaseparatio

nstep

shou

ldbe

integrated

into

theinstrument.

Specim

enprep

(totalstep

s)Integrated

specim

enprep

aration(includ

ingplasma

separatio

nifne

eded

);less

than

2step

srequ

ired(no

precisionvolumecontroland

precisiontim

estep

s)

Maxim

ally2step

s(noprecisionvolumecontroland

precisiontim

estep

s)Equipm

entsuchas

acentrifug

eor

heatblockareavailable

onlyinfrequ

ently

atlevel1

health

centresandsome

districth

ospitals,andthereforeshou

ldno

tberequ

iredfor

novelassays.Expertise

toop

erateaprecision

pipette

isalso

oftenlacking[21].

Forthe

detectionof

cAg,severalspecimen

preparation

stepsareneeded:i)todissociate

antib

ody-bo

undcAg;ii)

tolyse

viralparticlesandexpo

secAg;andiii)toinacti-

vate

antib

ody.Theseshou

ldalso

optim

allybe

integrated

with

thetestof

detection.

Timeto

result

<15

min

<60

min

Theneed

fora

rapidturn-aroun

dtim

e,thepo

ssibilityfor

batching

and/or

rand

omaccessfortesting,andthetest-

ingof

multiplespecimensatthesametim

eareinterre-

lated.Thetim

eto

resultisprob

ablythemostimpo

rtant

parameter,asextend

ingthewaittimeforp

atientswill

possiblyresultinlossto

follow-up[22,23].Mostcurrent

immune-chromatog

raph

icrapidtestsprod

uceresults

within20

min.

The Author(s) BMC Infectious Diseases 2017, 17(Suppl 1):707 Page 164 of 175

Table

1Com

bine

dHCVRN

AandcA

gTPPforan

HCVtest(Con

tinued)

Characteristic

Optim

alMinim

alRatio

naleandeviden

ce

Theidealtim

eto

resulthasno

tbe

enstud

iedand

might

vary

largelybe

tweencoun

triesandbe

tween

settings

whe

rethepatient

istested

.But

inorde

rto

bede

ployableas

atestforPO

C,the

resultshou

ldbe

availablewith

inthesamevisit.

Specim

encapacity

and

throug

hput

Multip

leat

atim

e;rand

omaccess/parallelp

rocessing

One

atatim

e(any

externalreagen

tsshou

ldbe

aliquo

tedforon

etim

euse)

Preferredthat

onespecim

endo

esno

toccupy

the

instrumen

tat

atim

e-i.e.,rand

omaccess/parallel

analysis.Iftheplatform

ismulti-analyte,then

runn

ing

different

assays

shou

ldbe

feasibleat

thesametim

e.

Biosafety+

waste

disposal

Mostly

simplewaste;m

inim

albiosafetywaste;no

sharps

None

edforabiosafetycabine

t;consum

ablesshou

ldbe

ableto

bedisposed

ofas

biosafetywaste;sim

ple

trash.

Increasedbiosafetyof

ano

veltestwillen

hance

acceptability

ofthetestby

providers.Furthe

rinform

ationprovided

inWHOLabo

ratory

Biosafety

Manual[24].

Instrumen

tatio

nInstrumen

t-fre

eAllow

forseparate

specim

enprep

arationde

vice

(e.g.

mini-cen

trifuge

)Thesim

pler,m

orepo

rtableanddu

rable/robu

stthetestis,

themorelikelyitwillbe

implem

entedinperipheral

setting

s.Ideally

aninstrumentfreetest(e.g.

immunochrom

atog

raph

ictest)w

ouldbe

thepreferred

optim

alsolutionbu

tthisislikelyno

tfeasib

lewith

the

analyticalsensitivitythatisnecessaryandasm

allspecimen

volumefro

mafingerstick.

Power

requ

iremen

tsIfde

vice

necessarythen

:battery-op

erated

with

recharging

solutio

n(e.g.solar)

andcircuitprotectorlastingup

to3days

ofconstant

useandableto

runoffstandard

electricity

Recharge

ablebatteryor

solarpo

wer

lastingat

least

8h.

Con

tinuo

uspo

wer

isno

talwaysavailableat

thelevel

ofahe

alth

andmicroscop

ycentre

andeven

less

likely

atprim

arycare

clinics,thereforeabattery-op

erated

de-

vice

with

charge

possibility

conceivablythroug

hsolar

power

wou

ldbe

mostidealinorde

rforatestto

fitinto

theen

tirebreadthof

settings

[21,23]

Mainten

ance/

calibratio

nDispo

sable,no

mainten

ance

orcalibratio

nrequ

ired

Ifde

vice

necessarythen

:preventativemainten

ance

at2yearsor

>5000

specim

ens;includ

emainten

ance

alert;remotecalibratio

n

Preven

tativemainten

ance

at1year

or>1000

specim

ens;on

lysimpletools/minim

alexpe

rtise

requ

ired;

includ

emainten

ance

alert.Sw

ap-out

ofplat-

form

spe

rmitted

.

Ifadevice

isanticipated

tohave

along

erlifespan,then

amaintenance

alertisessentialtoensure

prop

erfunctionalityinsettings

where

itisun

likelythatthesame

person

willalwayshand

lethede

vice

andrecordswillbe

kept

ondu

ratio

nof

use.

Itisessentialthato

nlysim

pletools/minimalexpertise

are

necessaryto

dothemaintenance

giventhatservicevisits

aredifficultou

tsideof

urbansetting

s.

Dataanalysis

Integrated

data

analysis

Integrated

data

analysis(norequ

iremen

tforPC

);expo

rted

data

capableof

beinganalysed

ona

separate

orne

tworkedPC

.

Con

nectivity

Ifde

vice

necessarythen

integrated

conn

ectivity;ifno

device

necessary,then

thetestshou

ldallow

data

expo

rtviaaseparate

reader.

Fulldata

expo

rt(onusageof

device,error/in

valid

rates,andpe

rson

alized

,protected

results

data)over

USB

portandne

twork.Networkconn

ectivity

throug

hEthe

rnet,W

iFi,and/or

GSM

/UMTS

mob

ilebroadb

and

mod

em.Resultsshou

ldbe

encode

dusinga

documen

tedstandard

(suchas

HL7)andbe

form

atted

asJSONtext.JSO

Ndata

shou

ldbe

transm

itted

Fulldata

expo

rt(onusageof

device,error/in

valid

rates,andpe

rson

alized

,protected

results

data)over

USB

portandne

twork.Networkconn

ectivity

throug

hEthe

rnet,W

iFi,and/or

GSM

/UMTS

mob

ilebroadb

and

mod

em.Resultsshou

ldbe

encode

dusinga

documen

tedstandard

(suchas

HL7)andbe

form

atted

asJSONtext.JSO

Ndata

shou

ldbe

transm

itted

throug

hhttp(s)to

alocalo

rremoteserver

asresults

arege

nerated.

Results

shou

ldbe

locally

stored

and

Dataexpo

rtwillen

hancesurveillance,de

vice

and

operator

managem

entandallow

forsupp

lychain

managem

ent.

The Author(s) BMC Infectious Diseases 2017, 17(Suppl 1):707 Page 165 of 175

Table

1Com

bine

dHCVRN

AandcA

gTPPforan

HCVtest(Con

tinued)

Characteristic

Optim

alMinim

alRatio

naleandeviden

ce

throug

hhttp(s)toalocalo

rremoteserver

asresults

arege

nerated.

Results

shou

ldbe

locally

stored

and

queued

durin

gne

tworkinterrup

tions

andsent

asa

batchwhe

nconn

ectivity

isrestored

.

queued

durin

gne

tworkinterrup

tions

andsent

asa

batchwhe

nconn

ectivity

isrestored

.

Resultcapture,

documen

tatio

n,data

display

Ifinstrumen

t-fre

e:ability

tosave

results

viaseparate

reader.

Ifde

vice

necessary:integrated

results

screen

and

ability

tosave

andprintresults;U

SBpo

rt.O

n-instrumen

tvisualreadou

tandtheability

toaddinfor-

mation(patient

ID,ope

ratorID,d

atelocatio

n,etc.)

Abilityto

save

results

Thetestmen

ushou

ldbe

simplewith

integrated

LCD

screen

;sim

plekeypador

touchscreen

.

Results

shou

ldbe

simpleto

interpret(positive/

negativeforHCVde

tection).

Ope

ratin

gtempe

rature/

humidity/altitude

Betw

een+5to

+40

oCat

90%

humidity

andat

analtitud

eof

3000

mBetw

een+10

oto

+35

oCat

70%

humidity

andat

analtitud

eof

2000

mHighenvironm

entaltem

peraturesandhigh

humidity

are

oftenaprob

lem

incoun

trieswhere

HCV

isendemic.

Reagen

tkit

transport

Nocold

chainrequ

ired;

toleranceof

transportstress

foraminim

umof

72hat

-15o

to+40

°CNocold

chainrequ

ired;

toleranceof

transportstress

foraminim

umof

48hat

-15o

to+40

oC

Refrige

ratedtransportiscostlyandoftencann

otbe

guaranteed

durin

gtheen

tiretransportatio

nprocess.

Freq

uent

delays

intransportarecommon

place.

Reagen

tkit

storage/stability

2yearsat

+5°C

to+40

oCat

90%

humidity

&transportstress

(72hat

50oC);no

cold

chainrequ

ired

12mon

thsat

+5°C

to35

oC,70%

humidity,including

transportstress

(48hat

50oC);no

cold

chainrequ

ired

Highenvironm

entaltem

peraturesandhigh

humidity

isoftenaprob

lem

inmanycountries

whereHCV

isprevalent.

Internalprocess

quality

control

Internalfull-processcontrol,po

sitivecontrol&

negative

controls

Externalpo

sitivecontrol

Inadditio

nto

compatib

ility

with

existin

gexternal

quality

assessmen

tsche

mes

Pricing

Maxim

umprice

forindividu

altest

(reagen

tcostson

ly;

atscale;ex-w

orks)

<US$5

<US$15

Fora

one-step

solutio

n,thecostneedsto

below,asa

trade-offintheease-of-u

se/perform

ance

forp

ricewou

ldno

tbe

accepted.Con

versely,inatwo-step

solutio

n,a

high

ercostismorelikelyto

beaccepted,aspeop

lewou

ldbe

willingto

makeatrade-offp

rovidedtheover-

allcostof

thealgo

rithm

remains

low.Cost-be

nefit

ana-

lysesareneeded

toexploredifferent

optio

ns.

Trade-offsbe

tweenop

timalcharacteristicsmay

bene

cessaryto

achieveop

timalpricing.

Preferen

ces

abou

tacceptabletrade-offsne

edto

befurthe

rde

fined

.

Maxim

umprice

forinstrumen

tatio

n<US$2000

<US$20,000

Thelower

thepriceforinstrumen

tatio

n,thelower

the

up-fron

tcostto

ahe

alth-caresystem

wou

ldbe

and

thus

thelower

the

barrierto

implem

entatio

n.Furthe

rmod

ellingisne

ces-

sary

toconfirm

themaxim

alpriceestim

ated

.Price

shou

ldinclud

ewarranties,servicecontractsandtech-

nicalsup

port.A

lternatively,rentalagreem

entsfor

equipm

entshou

ldbe

anop

tion.

TheTP

Pwas

finalized

usinginpu

tfrom

theDelph

i-likesurvey

anddiscussion

attheconsen

susmeetin

g(April22

,201

5,Vien

na,A

ustria).TP

Pswerecombine

dbe

causecharacteristicsweresimilar

forbo

th,ind

epen

dent

ofwhe

ther

thetest

envision

edwas

aRN

Aor

cAg-ba

sedtest.The

TPPne

edsto

beconsidered

inthecontextof

differen

ttype

sof

testingstrategies

(one

-step

versus

two-step

)*Com

paredto

aHCVRN

Areferencetest

performed

onplasma

**Ex-w

orks,including

prop

rietary

reag

ents

andconsum

ablescost

only

(with

outinstrumen

tatio

n),p

rodu

cedat

scale

The Author(s) BMC Infectious Diseases 2017, 17(Suppl 1):707 Page 166 of 175

The Author(s) BMC Infectious Diseases 2017, 17(Suppl 1):707 Page 167 of 175

to 5 (1 = disagree, 2 = mostly disagree, 3 = do notagree or disagree, 4 = mostly agree, 5 = fully agree).Consensus was pre-defined as more than 50% of re-spondents providing a score of at least four on theLikert scale (as per WHO policy for expert group con-sensus). Initially, one round was planned followed by ananonymous summary of the experts’ opinion, followedagain by a second survey round. However, since prede-fined consensus for all characteristics was reached afterthe first round, a second round was not initiated andthus a “Delphi-like” method was used.

5. A final consensus-gathering meeting for further dis-cussions around disputed characteristics. Aconsensus-gathering meeting was convened by FINDand the Forum, in collaboration with WHO, onApril 22, 2015, prior to the European Associationfor the Study of the Liver’s annual meeting inVienna, Austria. The invitation to the meeting wassent to 50 organizations/individuals who participatedin the TPP survey (step 4) and 35 people were ableto attend. No financial support was provided to par-ticipants at the meeting. Given the high level of con-sensus in the Delphi-like process, the main purposeof this meeting was to discuss issues on which fewerthan 75% of the respondents agreed, or on which adistinct subgroup disagreed, or on which furtherneed for clarification and discussion was considereduseful. If any voting was necessary throughout theconsensus-gathering meeting, >75% was considereda majority.

Values and preferences survey of implementers and usersof hepatitis C testsIn the development of the WHO 2017 guidelines on hepa-titis B and C testing, the GRADE system (Grading of Rec-ommendations, Assessment, Development and Evaluation)was used to categorise both the strength of recommenda-tions as strong or conditional (based on consideration ofthe quality of evidence, balance of benefits and harms, ac-ceptability or values and preferences of end-users, resourceuse and programmatic feasibility). To assess the values andpreferences for trade-offs between different test characteris-tics in the TPP, different testing strategies and approaches,a four-part online survey tool was developed by FIND andWHO with input from MSF, CDC and Médecins duMonde using SurveyMonkey software (SurveyMonkey Inc.,Palo Alto, California, USA) Ivanova et al [35]. Survey ques-tions were designed as multiple-choice with commentssections where appropriate (see Additional file 1; only thequestions relevant to further the understanding around userpreferences in regards to the TPP for HCV diagnostics arelisted). The survey was undertaken in September 2015, andthe invitation to participate was sent via email to 306 peopleon the WHO Hepatitis database and included in the FIND

and Forum newsletters, and included clinicians, patient orga-nizations, civil society representatives, national hepatitisprogramme managers, policy-makers and diagnostics andpharmaceutical industry employees. Responses wereexported into a Microsoft Excel table (Microsoft Office2010, Microsoft Corporation). A descriptive analysis wasperformed using R statistical software (version 3.2, R Foun-dation for Statistical Computing). Responders were not re-quired to answer to all questions, thus the denominator foreach response are the number of responders to the specificquestion. Only findings relevant to user preferences inregards to the TPP for HCV diagnostics are presented here.

ResultsStakeholder participants in TPP developmentOf the 50 organizations/individuals invited to participate inthe Delphi-like TPP consultation, 36 responded to the HCVRNATPP and 26 to the cAg TPP, and 35 attended the con-sensus meeting in Vienna. About half of responders werefrom the in vitro diagnostics industry or product develop-ment partnerships/technical agencies/researchers (27% and25% respectively), 14% were from advocacy organizationsand the same from the pharmaceutical industry, 11% wereimplementers/clinicians and the remainder (3% each) repre-sented national hepatitis programmes, international bodiesand consultants (Fig. 2).

Consensus on TPP characteristicsThe consensus process was used to define the character-istics within the four categories of the TPP: scope, per-formance, operational characteristics and price. Of the27 characteristics across the four categories, there wasgreater than 50% agreement on all of the TPP character-istics. There was even greater than 75% agreement forall characteristics except the minimum clinical sensitivityand the price. All stakeholders agreed that the TPP forHCV RNA and for HCV cAg test could be combinedinto one TPP for a near-patient HCV virological test.

Target specifications for key characteristics of TPP forHCV virological testThe final characteristics in the TPP are summarized inTable 1. Key characteristics are described below.

Scope of the test

Goal of the test The near-patient test defined in the TPPis for diagnosis of active, viraemic HCV infection. Two test-ing strategies are considered: (i) a one-step strategy using asingle HCV RNA or cAg test (ii) a two-step strategy usingan HCV antibody screening test followed by a confirmatoryHCV RNA or cAg test, if antibody positive (Fig. 1). Con-ceivably, provided the prevalence is substantial and the costof the test is sufficiently low, it could be used in a one-step

Fig. 2 Information about respondents. Professional profiles of 36 respondents to the HCV RNA and cAg TPPs

The Author(s) BMC Infectious Diseases 2017, 17(Suppl 1):707 Page 168 of 175

diagnostic strategy, thus obviating the need for an upfrontantibody screening test. Additionally, since the risk of re-infection is substantial in some population groups, thismeans that certain high risk groups who have been treatedfor HCV infection in the past should be tested again forHCV RNA or HCV cAg directly as HCV antibody screen-ing test is expected to be positive.Furthermore, the test should be also usable to confirm

cure upon treatment completion.

Target population The target population for the testwas considered to be all persons in countries with amedium to high prevalence of HCV (1.5–3.5% and>3.5%) as well as high-risk populations in low prevalencesettings (<1.5%).

Target operator The target operator of a test optimallywas considered to be a community worker with minimaltraining particularly in the context of a one-step testingstrategy, where easy-to-use testing close to the patient isparamount to increase access to care. At a minimum thetest should be easy enough to perform by a health-careworkers or laboratory technicians with limited training(i.e. the test would need to be integrated test with min-imal additional and no precision steps; see operationalcharacteristics below).

Performance characteristics

Analytical and clinical sensitivity The analytical sensi-tivity or the lower limit of detection (LOD) describe the

The Author(s) BMC Infectious Diseases 2017, 17(Suppl 1):707 Page 169 of 175

smallest concentration of an analyte that can be reliablymeasured by an analytical procedure. Current availableHCV RNA assays are able to detect as low as 5 IU/ml[18]. Trade-offs such as cost and ease-of use have to beconsidered in setting targets for analytical sensitivity. Ifthe optimal specimen type (i.e. capillary blood; see below)or a simplified transport tool (e.g. dried blood spots, DBS)were considered, a lower analytical sensitivity (<1000 IU/ml) would be technically difficult to achieve.In the proposed TPP, the optimal analytical sensitivity

was defined as 200 IU/mL for both HCV RNA and cAgtests, while the minimum acceptable level was 1000–3000 IU/mL. Translating the proposed minimal analyticalsensitivity into clinical sensitivity (i.e. the probability thatthe test will correctly identify an infected individual), theLOD would result in >95% of viraemic patients correctlydiagnosed based on available limited data on distributionof viral load measurements in those diagnosed with vir-aemic HCV infection [29, 30]. The majority of stake-holders felt that from a population perspective any testwith 95% sensitivity that would substantially improve ac-cess to diagnosis and treatment would be a step forward.Further, it was suggested that an optimal sensitivity of200 IU/ml would likely be sufficient to detect >99% of vir-aemic patients. Both limits of detection need to be vali-dated in larger surveillance studies on quantitative viralload measurement.

Polyvalency (i.e., same device can be used for differenttests) Polyvalency was considered a key characteristic toincrease affordability of and access to a diagnostic testby the stakeholder group. Given that polyvalency wasinitially not voted upon in the Delphi-like survey, a votewas taken in the consensus-meeting. More than three-quarters of stakeholders supported the development ofpolyvalent platforms that integrate diagnosis of multiplediseases such as HIV, TB and hepatitis B with HCV asan optimal characteristic. Multiplexing (i.e. detectingdifferent analytes in a single sample at the same time)was seen as advantageous for some tests by but not con-sidered as essential by the majority.

Quantitation The majority of stakeholders (>85%)thought that a quantitative test is no longer required ifDAA regimens are used, as a decision to treat and con-firmation of cure is based on presence of viraemia, andresponse-guided therapy is likely no longer required orfeasible [31]. However, several stakeholders highlightedthat a quantitative test is still advantageous to addressspecific research questions. Furthermore, diagnosticmanufacturers suggested that there is only a small costdifferential for the development of a qualitative andquantitative test. Therefore, quantitation was integratedas an optimal target.

Operational characteristics

Specimen type All stakeholders (100%) agreed that theoptimal test is the one that can be performed on capil-lary whole blood without any requirement for any add-itional laboratory equipment as the need for phlebotomyto draw venous whole blood would limit the applicabilityof the test in lower settings of care. If plasma is to be aspecimen type (minimal criteria), the plasma separationstep should be integrated into the instrument.

Specimen preparation For the specimen preparation,the majority (>86%) of stakeholders suggested that an in-tegrated specimen preparation (including plasma separ-ation if needed) would be optimal. At a minimum, lessthan 2 steps should be required (with no steps beingprecision volume control and precision time steps).These requirements are in line with the skills set a testoperator would be expected to have at the settings ofintended use [32].

Time to result/throughput The optimal turnaroundtime defined in the TPP is less than 15 minutes and, at aminimum, the test results should be delivered within1 hour. The need for a rapid turn-around time, the possi-bility for batching and/or random access for testing, andthe testing of multiple specimens at the same time are in-terrelated. The time to result is probably the most import-ant parameter, as extending the wait time for patients willpossibly result in loss to follow-up [33, 34]. The ideal timeto result might vary largely between settings where the pa-tient is tested and has been further evaluated in the valuesand preferences survey. In order to be deployable as a testfor point-of-care, the result should be available within thesame visit. Most current immune-chromatographic rapidtests produce results within 20 minutes.

Instrumentation and power requirements The idealtest was considered to be an instrument-free test. If aninstrument is necessary, then it needs to be battery oper-ated, ideally with a recharging solution (e.g. solar) andcircuit protector lasting up to 3 days of constant use.For the description of other performance and oper-

ational characteristics please refer to the Table 1.

Price of the testThe specification on the ex-works test price (i.e. price oftest at the manufacturing factory) at scale had the lowestlevel of agreement among stakeholders (only 58% agree-ment for minimal price for HCV RNA test at <US $ 15and 73% for HCV cAg test at <US $ 10), while 75% agreedon an optimal cost of an HCV RNA of <7 $ US and 85%agreed on an optimal cost of a HCV cAg test at <US $ 3.The opinions on the minimal cost were split between

The Author(s) BMC Infectious Diseases 2017, 17(Suppl 1):707 Page 170 of 175

stakeholders, with many considering the price was too lowand others the price was too high. During the discussion,the main factors raised by the stakeholders that would in-fluence the price of tests included the following: (i) thecomplexity of the test, for example, if there was complexspecimen preparation required for the release and detec-tion of HCV cAg; and (ii) the volume of tests undertaken,which was expected to be lower initially given the antici-pated lower demand for HCV diagnostics, as compared todiagnostics for other diseases such as HIV due to the lackof hepatitis C care services in LMICs. Furthermore, stake-holders highlighted that the ex-works price, while bettercompared to other diagnostics, would not be the priceseen by the end-user as shipping costs, import taxes, cus-toms charges and local distribution costs would be addedon. The control and limitation of these add-on costs wasconsidered paramount to enable the scale up of testing.Ultimately the cost was set at US $ 5 optimally and US $15 minimally and the majority of stakeholders (>75%)agreed under the precondition that substantial trade-offswould have to be accepted (e.g. in regards to sensitivity)with a test at US $ 5.

Values and preferences (V&P) surveyThe V&P survey questionnaire covered current hepatitisB and C testing practices and preferences for future test-ing diagnostics. Only the findings relevant to further theunderstanding around user preferences in regards to theTPP for HCV diagnostics are presented here (seeAdditional file 1), while the major part of the survey re-sults is reported separately [35].In total, 104 respondents from 43 countries (56 re-

spondents from 20 high-income countries and 48 re-spondents from 23 LMIC) participated in the survey.The absolute majority of respondents (94 out of 104)were implementers (i.e. people working in projects forHCV care in countries either within governmental ornon-governmental organizations) and IVD industry em-ployees accounted for only 5.8% (6/104). More than ahalf of the respondents (65 out of 104) had more than10 years of experience in the field of viral hepatitis.Responses to survey questions on several key TPP

characteristics were as follows:Test of cure: Almost half of the respondents (47 out of

99) considered that the diagnostic test for HCV viraemicdiagnosis should the same as the one for test of cure(Fig. 3). Of these forty-seven, 90% preferred a HCV RNAtest (decentralized or laboratory-based) primarily be-cause of its high accuracy over a cAg test.To explore preferences for future HCV diagnostics in

resource-limited settings, the survey specifically col-lected opinions of stakeholders from LMIC on desiredperformance and operational characteristics of futureHCV test and acceptable trade-offs.

Sensitivity: Half of the respondents (23 out of 46 LMICrespondents) considered that 95% was the lowest accept-able level of clinical sensitivity (Fig. 4b), and 43% (20 outof 46) preferred a sensitivity of 98%. However, when askedabout the price for tests with different sensitivity, most ofthe respondents (83%) indicated that a test with 95% sen-sitivity is acceptable if the price is lower than US $ 10(Fig. 4c and d). Only 7% (3 out of 46) felt that a 95% sensi-tivity would be unacceptably low.Specimen type: Half of respondents (21 out of 45)

expressed a preference for a test that can be performed oncapillary blood even if this was at the expense of sensitivity(Fig. 5a). The majority (75%) also considered the ability ofthe test to use dried blood spots (DBS) as important or veryimportant particularly for centralized testing (Fig. 5b).Turnaround time: 21% of respondents considered

30 minutes the optimal turnaround time, while 14% and16% were willing to accept 1 or 2 hours, respectively.Twenty-seven percent were willing to accept a turn-around time of 1 day and 14% allowed even for the nextday (Fig. 6).Price: The majority of respondents considered current

prices of virological tests as a barrier to implementation,particularly for HCV RNA tests (77%) and less so forcAg tests (60%) (Fig. 4a). When asked about a maximumacceptable price per test (ex-works and at scale), 52%(24 out of 46) of respondents indicated US $ 1–10 for atest with 98% sensitivity, while 41% (19 out of 46) wereready to pay up to US $ 20 (Fig. 4c). Half of respondents(21 out of 46) would also accept a 95% sensitive test atthe cost of US $ 1–10, while 37% (17 out of 46) consid-ered that such a test should cost < US $ 1 (Fig. 4d).Testing strategy: When asked about preferred testing

strategy (Fig. 1), half of respondents (50%) considered atwo-step approach with an HCV serological test followedby confirmatory HCV virological test to be an optimaltesting strategy. Almost the same number of respondents(48%) expressed preference for a one-step testing strategythat implies a single HCV virological POC. They explainedtheir preference by the simplicity of the algorithm and re-duced rate of loss-to-follow-up (Fig. 7a). If a one-stepstrategy is used, 52% of the respondents (24 out of 46)preferred an HCV RNA test as a single virological test,while 35% (16 out of 46) considered that an HCV cAg testwould be an optimal diagnostic solution (Fig. 7b).

Discussion and conclusionsFor the next generation of near-patient/point-of-carediagnostic tests for HCV viraemia based on HCV RNAor HCV cAg detection to succeed in improving access tocare, certain key specifications need to be met. We re-port on these key specifications for scope, performance,operational characteristics and price in a TPP utilizing aconsensus-based process with key global stakeholders.

Fig. 3 Preference for the HCV diagnostic test and the test of cure to be the same or different. Almost half of respondents preferred that thediagnostic test and test of cure be the same and about two thirds of these preferred that it be a decentralized (i.e. near-patient) HCV RNA test

The Author(s) BMC Infectious Diseases 2017, 17(Suppl 1):707 Page 171 of 175

The most critical features of the assay identified are sen-sitivity, specimen type, steps required to perform the testand price. The additional V&P survey that was per-formed to understand the values and preferences of endusers and implementers around specifications set in theTPP, highlights potential compromises on performancefeatures to achieve lower cost and ease-of-use.The TPP sets high requirements for optimal test ac-

curacy (>99%) but set a lower target under minimal cri-teria (90–95%). The V&P survey showed that a majority

Fig. 4 Preferred sensitivity and acceptable trade-offs between sensitivity anRNA, and to a lesser extent HCV cAg, is seen as a major barrier to scale updiagnostic sensitivity for a one-step diagnostic test whereas 43% prefer 98%were willing to compromise on sensitivity for a cheap test c and d

of end-users and implementers considered a test with95% sensitivity to be acceptable while 90% sensitivitywas considered to be insufficient. Modelling studies arerequired in order to determine the minimum sensitivitythat is needed for the point-of-care test to lead to thesubstantial improvement in HCV detection on a popula-tion level. Also, only limited data are available to under-stand the limit of detection that corresponds to 90–95%clinical sensitivity. Available data show that the majorityof infected individuals with chronic HCV infection have

d price. Over 50% of respondents see current price of HBV DNA, HCVthe testing a. Half of respondents consider that 95% as an acceptableb. Respondents were willing to pay more for a higher sensitivity and

Fig. 5 The importance of specimen type and the ability to use dried blood spots. Capillary and venous blood are the preferred specimen typesa and most respondents consider the ability to use DBS as a specimen type important or very important b

The Author(s) BMC Infectious Diseases 2017, 17(Suppl 1):707 Page 172 of 175

viral loads between 104 and 107 IU/mL, however a tran-sient dip to values below 103 IU/ml can be observed inpatients with partial viral control within the first6 months [29]. Thus, the optimal limit of detection of200 IU/mL set in the proposed TPP should detect al-most all viraemic patients (i.e. >99%(LOD) i.e. optimal).However, the data are limited in terms of capturing dif-ferent populations, stages of disease and genotypes andfurther comprehensive studies are needed to confirm theacceptable LOD. The V&P survey indicated that manystakeholders would be willing to compromise on sensitiv-ity if such a test could be achieved at decreased cost andincreased ease-of-use. Certainly, test design decisions suchas the use of capillary blood for a specimen will limit theability to achieve a LOD below 1000 IU/mL. The detec-tion of cAg instead of RNA will also limit the ability toreach an optimal analytical sensitivity, as even laboratory-based assays that include complex specimen preparation

Fig. 6 Acceptable turn around times between taking the specimen and reresult is preferred

(i.e. Abbott Architect) can detect femtomolar amounts ofcAg and have a LOD equivalent to about 3000 IU/ml foran RNA assay.The TPP also sets challenging operational characteristics

to ensure the usability of the test in low levels of thehealthcare system, close to where the patients firstpresent. Capillary blood is defined as an optimal specimentype as it greatly reduces the need for skilled workers, pre-cision pipetting, centrifugation or nearby complex labora-tory infrastructure, which is particularly beneficial atlower levels of the healthcare system where these re-sources and skills are lacking [32, 34]. The acceptable andoptimal turnaround time will enable diagnostic decisionmaking possibly on the same day, reduce costs of trans-portation for patients and so in turn loss to follow-up[33]. However, results of the V&P survey suggest that aneven longer turn-around time would be acceptable insome settings. Other operational characteristics, such as

turning the result to the service provider. A maximum of a same day

Fig. 7 Preferences for HCV testing strategy. Half of respondents consider two-step testing strategy (see Figure 1) to be an optimal future testingapproach in LMIC, while the other half prefer currently used two-step approach a. If one-step testing strategy is implemented, more than a halfof respondents prefer to use HCV RNA test b

The Author(s) BMC Infectious Diseases 2017, 17(Suppl 1):707 Page 173 of 175

the requirement for an instrument to run of batterypower, the need for stability in wide ranges of temperatureand humidity reflect upon the possible settings the testcould be used in LMIC.A further consensus conclusion is that on-treatment

monitoring does not appear to be necessary or usefulwith DAA regimens as the early reduction in viral loaddoes not correlate with cure [31]. However a test of cureis still important. Given the high correlation betweenHCV cAg and HCV RNA levels [36], either test wouldlikely be suitable as a test-of-cure. Again, limited data isavailable on the required analytical sensitivity for a testof cure. Cohn et al. have suggested that 3000 IU/mL islikely suffice for this indication [37], however furtherdata are needed about the rapidity and trajectory of aviral load rebound after treatment failure across differentpopulations, fibrosis stages and genotypes particularlywith new DAA regimens. The WHO guidelines recom-mend a single test of cure performed either at 12 weeksor 24 weeks after the end of treatment [1].The topic on which unsurprisingly there was the least

consensus between stakeholders in the meeting as wellas in the V&P survey was the price per test, with end-users and manufactures disagreeing on acceptable prices.A key conclusion was the need for greater transparency,both from the manufacturers themselves on cost ofgoods and volume-based price-breaks, but also from im-plementers on cost of distribution so that costs can bedisaggregated and end-users can gauge what a fair priceshould be for the goods purchased. Approaches to redu-cing costs include (i) countries and donors coming to-gether to pool procurement and forecasting; (ii) launchof competitive tenders; (iii) request bundled pricing onthe back of HIV and TB test roll-out; (iv) exploration ofdifferent supply chain models for LMIC to reduce thedistribution costs incurred; and (v) integrate HCV carewithin programs for other diseases, such as HIV, TB and

hepatitis B, to leverage investments made in polyvalentplatforms and care infrastructure. Particularly for thepooled procurement and forecasting there are many les-sons learned from the HIV experience.The strengths of this work include:- (i) it followed a

formalized step-wise process established for the develop-ment of consensus-based TPPs prior; (ii) It included abroad group of stakeholders; (iii) the TPP process wascomplementary by a survey that was able to elucidatetrade-offs between characteristics. However, we also haveto acknowledge limitations. The TPP reflects the opinionof the stakeholders represented in the Delphi-like surveyand at the meeting. While we attempted to have a largegroup of stakeholders that was representative of all thedifferent groups (national Hepatitis programs; imple-menters and clinicians (including from NGOs); technicalagencies and researchers; international organizations;advocacy organizations/civil society; industry (pharma-ceutical and IVD-manufacturers), not everybody was ableto participate. It also has to be noted that industry repre-sentatives accounted for a large percentage of the stake-holders however an analysis of the TPP survey excludingthe responses from industry representatives showed min-imal differences to an overall analysis (data not shown).As part of the discussion the following key research

questions were identified, that should be addressed toinform future test development and guidelines:

1. Improve understanding on what clinical sensitivitywould be achieved with an optimal analyticalsensitivity of 200 IU/ml and a minimum analyticalsensitivity of 1000–3000 IU/ml. Also, what are thecharacteristics of HCV-infected individuals with aviraemia of <1000–3000 IU/ml who are missed byan assay with minimum sensitivity? Are they lessprone to develop substantial HCV disease? Are theymore likely to resolve their infection?

The Author(s) BMC Infectious Diseases 2017, 17(Suppl 1):707 Page 174 of 175

2. What should be the sensitivity of a test of cure andwhen should such a test be best performed tooptimize feasibility, ideally with the same platform asused for detection and minimize loss-to-follow up.

3. What is the cost effectiveness of a one-step approachversus a conventional two-step approach (with HCVserological test first followed by HCV RNA or cAgtest) in different prevalence settings?

The ability to diagnose HCV infection in LMIC near tothe patient or at the point of care is an important aspectof scaling up access to HCV care. While centralized plat-forms are already available, an accurate, easy-to-use andaffordable diagnostic solution to confirm active HCV in-fection near to the patient is lacking and urgently neededin LMIC. Independently, such a test would likely be of usein expanding diagnostic testing in resource-rich settings.In considering the needs in resource limited settings, thestakeholder community and donors need to come to-gether to now translate this high priority TPP into actualtests that can help to curb the HCV epidemic and ensureaffordable and accessible diagnosis for HCV.

Additional file

Additional file 1: The V&P survey questionnaire. One table in MicrosoftWord format that where survey questions are listed in the left column andanswer options in the right. Only the questions linked to data presentedhere are listed. (DOCX 21 kb)

AcknowledgementsGrateful thanks to the late Dr. Samuel Lattimore.

FundingPublication of this article was funded by the World Health Organization.

Availability of data and materialsNot Applicable.

About this supplementThis article has been published as part of BMC Infectious Diseases Volume 17Supplement 1, 2017: Testing for chronic hepatitis B and C – a globalperspective. The full contents of the supplement are available online athttps://bmcinfectdis.biomedcentral.com/articles/supplements/volume-17-supplement-1.

Authors’ contributionsTPP development: Claudia Denkinger, Philippa Easterbrook, AlessandraTrianni, Nivedha Panneer, Veronica Miller; The V&P survey: Elena IvanovaReipold, Douglas Krakower, Stefano Ongarello, Teri Roberts; Drafting ofmanuscript: Claudia Denkinger, Philippa Easterbrook, Elena Ivanova, TeriRoberts; Critical revision of manuscript: Claudia Denkinger, PhilippaEasterbrook; All authors read and approved the final manuscript.

Ethics approval and consent to participateNot Applicable.

Consent for publicationNot Applicable.

Competing interestsThe authors declare that they have no competing interests.

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

Author details1FIND, MSF Access Campaign, Geneva, Switzerland. 2Forum for CollaborativeHIV Research, Washington D.C, USA. 3Beth Israel Deaconess Medical Center,Harvard Medical School, Boston, MA, USA. 4Global Hepatitis Programme, HIVDepartment, World Health Organization, Geneva, Switzerland.

Published: 1 November 2017

References1. Guidelines on hepatitis B and C testing. Geneva: World Health Organization;

2017. http://apps.who.int/iris/bitstream/10665/254621/1/9789241549981-eng.pdf?ua=1. Accessed June 2017.

2. Mohd Hanafiah K, Groeger J, Flaxman AD, Wiersma ST. Global epidemiologyof hepatitis C virus infection: New estimates of age-specific antibody to HCVseroprevalence. Hepatology. 2013;57:1333–42.

3. Global Hepatitis Report 2017. Geneva: World Health Organization; 2017.http://apps.who.int/iris/bitstream/10665/255016/1/9789241565455-eng.pdf?ua=1. Accessed June 2017.

4. Pol S, Corouge M, Vallet-Pichard A. Daclatasvir-sofosbuvir combinationtherapy with or without ribavirin for hepatitis C virus infection: from theclinical trials to real life. Hepat Med. 2016;8:21–6.

5. Andrieux-Meyer I, Cohn J. Affonso de Araujo E, Hamid S. Disparity in marketprices for hepatitis C virus direct-acting drugs. Lancet. 2015;3:e676–7.

6. Gilead. Chronic Hepatitis C Treatment Expansion Generic Manufacturing forDeveloping Countries. August. 2015. p. 2. http://www.gilead.com/~/media/Files/pdfs/other/HCVGenericAgreementFactSheet.pdf. Accessed June 2017.

7. The Medicines Patent Pool Signs First Sub-licences for Hepatitis C MedicineDaclatasvir. Available from https://medicinespatentpool.org/mpp-media-post/the-medicines-patent-pool-signs-licence-with-bristol-myers-squibb-to-increase-access-to-hepatitis-c-medicine-daclatasvir/. 2016. Accessed June 2017.

8. Hill A, Bryony S, Dzintars G, Joseph F. Rapid reductions in prices forgeneric sofosbuvir and daclatasvir to treat hepatitis C Rapid reductionsin prices for generic sofosbuvir and daclatasvir to treat hepatitis C. JVirus Erad. 2016;2:28–31.

9. Hill A, Khoo S, Fortunak J, Simmons B, Ford N. Minimum costs for producingHepatitis C Direct Acting Antivirals, for use in large-scale treatment accessprograms in developing countries. Clin Infect Dis. 2015;61:1174–82.

10. Denkinger CM, Kessel M. Diagnostics for hepatitis C: an urgent need foraction. Lancet Glob Heal. Denkinger et al. Open access article distributedunder the terms of CC BY; 2015;3:e195.

11. Vos T, Barber RM, Bell B, Bertozzi-Villa A, Biryukov SBI, et al. Global, regional,and national incidence, prevalence, and years lived with disability for 301acute and chronic diseases and injuries in 188 countries, 1990–2013: asystematic analysis for the Global Burden of Disease Study 2013. Lancet.2015;386:743–800.

12. Murtagh M. UNITAID Hepatitis C Diagnostics Technology Landscape 1stEdition. Geneva: 2015. http://unitaid.org/images/marketdynamics/publications/UNITAID-HCV_Diagnostic_Landscape-1st_edition.pdf.Accessed June 2017.

13. Putting HIV and HCV to the Test: A Product Guide for Point-of-Care CD4and Laboratory-Based and Point-of-Care Virological HIV and HCV Tests.Geneva: Médecins Sans Frontières (MSF); 2015:1–100. http://www.msfaccess.org/sites/default/files/HIV_HCV_Report_Diagnostic_Guide_2015.pdf.Accessed Nov 2016.

14. Roberts T, Cohn J, Bonner K, Hargreaves S. Scale-up of Routine Viral LoadTesting in Resource-Poor Settings: Current and Future ImplementationChallenges. Clin Infect Dis. 2016;62:1043–8.

15. Peterson J, Green G, Lida K, Lee S. Detection of Hepatitis C Core Antigen inthe Antibody Negative “Window” Phase of Hepatitis C Infection. Vox Sang.2000;78:80–5.

16. Chevaliez S, Feld J, Cheng K, Wedemeyer H, Sarrazin C. Clinical utility ofHepatitis C virus core antigen detection and quantification in the diagnosisand management of patients with chronic hepatitis C receiving an all-oral,interferon-free regimen. Antivir Ther. 2016. doi:10.3851/IMP3042. [Epubahead of print]

17. Medici MC, Furlini G, Rodella A, Fuertes A, Monachetti A, Calderaro A, et al.Hepatitis C virus core antigen: Analytical performances, correlation with

The Author(s) BMC Infectious Diseases 2017, 17(Suppl 1):707 Page 175 of 175

viremia and potential applications of a quantitative, automatedimmunoassay. J Clin Virol. 2011;51:264–9.

18. Cepheid, Sunnyvale U. Cepheid brochure: Xpert HCV Viral load. http://www.cepheid.com/administrator/components/com_productcatalog/library-files/611eae732b51634d8a7e64878d8222a1-615085cee2245c0c5e547ebff5052119-Xpert-HCV-Viral-Load-Brochure-CEIVD-3043-02.pdf.Accessed Oct 2016.

19. Price CP. Point-of-care testing. Impact on medical outcomes. Clin Lab Med.2001;21:285–303.

20. Target product profiles and priority digital health products for TB. Geneva:World Health Organization; 2015. http://www.who.int/tb/areas-of-work/digital-health/target-product-profiles/en/. Accessed June 2017.

21. High-priority target product profiles for new tuberculosis diagnostics: reportof a consensus meeting. Geneva: World Health Organization; 2014. http://apps.who.int/iris/bitstream/10665/135617/1/WHO_HTM_TB_2014.18_eng.pdf?ua=1&ua=1. Accessed June 2017.

22. U.S. Department of Health and Human Services, Food and DrugAdministration, Center for Drug Evaluation and Research (CDER). Guidancefor Industry and Review Staff: Target Product Profile — A StrategicDevelopment Process Tool. Rockville MD. 2007. Available from http://www.fda.gov/downloads/drugs/guidancecomplianceregulatoryinformation/guidances/ucm080593.pdf. Accessed June 2017.

23. STI POCT Meeting report: Point of Care Tests –Target Product Profiles andResearch Questions, May 6–8, Annecy, France. http://who.int/reproductivehealth/POTC-TPPs-2016.pdf?ua=1. Accessed June 2017.

24. Denkinger CM, Dolinger D, Schito M, Wells W, Cobelens F, Pai M, et al.Target product profile of a molecular drug-susceptibility test for use inmicroscopy centers. J Infect Dis. 2015;211:S39–49.

25. Porrás AI, Yadon ZE, Altcheh J, Britto C, Chaves GC, Flevaud L, et al. TargetProduct Profile (TPP) for Chagas Disease Point-of-Care Diagnosis andAssessment of Response to Treatment. Debrabant A, editor. PLoS Negl TropDis Public Libr Sci; 2015;9:e0003697.

26. Price CP. Medical and Economic Outcomes of Point-of-Care Testing. ClinChem Lab Med. 2002;40:246–51.

27. Hyle EP, Jani I V., Lehe J, Su AE, Wood R, Quevedo J, et al. The Clinical andEconomic Impact of Point-of-Care CD4 Testing in Mozambique and OtherResource-Limited Settings: A Cost-Effectiveness Analysis. Sansom SL, editor.PLoS Med Public Libr Sci; 2014;11:e1001725.

28. Rowe G, Wright G. The Delphi technique as a forecasting tool: issues andanalysis. Int J Forecast. 1999;15:353–75.

29. Glynn SA, Wright DJ, Kleinman SH, Hirschkorn D, Tu Y, Heldebrant C, et al.Dynamics of viremia in early hepatitis C virus infection. Transfusion. 2005;45:994–1002.

30. Fytili P, Tiemann C, Wang C, Schulz S, Schaffer S, Manns MP, et al.Frequency of very low HCV viremia detected by a highly sensitive HCV-RNAassay. J Clin Virol. 2007;39:308–11.

31. Sidharthan S, Kohli A, Sims Z, Nelson A, Osinusi A, Masur H, et al. Utility ofHepatitis C Viral Load Monitoring on Direct-Acting Antiviral Therapy. ClinInfect Dis. 2015;60:1743–51.

32. Denkinger C, Nicolau I, Ramsay A, Chedore P, Pai M. Are peripheralmicroscopy centres ready for next generation molecular tuberculosisdiagnostics? Eur Respir J. 2013;42:544–7.

33. Sreeramareddy CT, Kishore PV, Menten J, Van den Ende J. Time delays indiagnosis of pulmonary tuberculosis: a systematic review of literature. BMCInfect Dis. 2009;9:91.

34. Usdin M, Guillerm M, Calmy A. Patient needs and point-of-carerequirements for HIV load testing in resource-limited settings. J Infect Dis.2010;201:S73–7.

35. Ivanova E, Trianni A, Krakower D, Ongarello S, Roberts T, Easterbrook P andDenkinger C. Values, preferences and current hepatitis B and C testingpractices in low- and middle-income countries: results of a survey of end usersand implementers. BMC Infect Dis. 2017;17:(Suppl 1). doi:10.1186/s12879-017-2769-y.

36. Hadziyannis E, Minopetrou M, Georgiou A, Spanou F, Koskinas J. Is HCV coreantigen a reliable marker of viral load? An evaluation of HCV core antigenautomated immunoassay. Ann Gastroenterol. 2013;26:146–9.

37. Cohn J, Roberts T, Amorosa V, Lemoine M, Hill A. Simplified diagnosticmonitoring for Hepatitis C, in the era of Direct Acting Antiviral treatment.Curr Opin HIV AIDS. 2015;10:369–73.

• We accept pre-submission inquiries

• Our selector tool helps you to find the most relevant journal

• We provide round the clock customer support

• Convenient online submission

• Thorough peer review

• Inclusion in PubMed and all major indexing services

• Maximum visibility for your research

Submit your manuscript atwww.biomedcentral.com/submit

Submit your next manuscript to BioMed Central and we will help you at every step: