Target product profiles for tests

for tuberculosis treatment

monitoring and optimization
Target product profiles for tests
for tuberculosis treatment
monitoring and optimization
Target product profiles for tests for tuberculosis treatment monitoring and optimization

ISBN 978-92-4-008117-8 (electronic version)

ISBN 978-92-4-008118-5 (print version)

© World Health Organization 2023

Some rights reserved. This work is available under the Creative Commons Attribution-NonCommercial-ShareAlike 3.0 IGO
licence (CC BY-NC-SA 3.0 IGO; https://creativecommons.org/licenses/by-nc-sa/3.0/igo).

Under the terms of this licence, you may copy, redistribute and adapt the work for non-commercial purposes, provided the
work is appropriately cited, as indicated below. In any use of this work, there should be no suggestion that WHO endorses
any specific organization, products or services. The use of the WHO logo is not permitted. If you adapt the work, then you
must license your work under the same or equivalent Creative Commons licence. If you create a translation of this work,
you should add the following disclaimer along with the suggested citation: “This translation was not created by the World
Health Organization (WHO). WHO is not responsible for the content or accuracy of this translation. The original English
edition shall be the binding and authentic edition”.

Any mediation relating to disputes arising under the licence shall be conducted in accordance with the mediation rules of
the World Intellectual Property Organization (http://www.wipo.int/amc/en/mediation/rules/).

Suggested citation. Target product profiles for tests for tuberculosis treatment monitoring and optimization. Geneva:
World Health Organization; 2023. Licence: CC BY-NC-SA 3.0 IGO.

Cataloguing-in-Publication (CIP) data. CIP data are available at http://apps.who.int/iris.

Sales, rights and licensing. To purchase WHO publications, see https://www.who.int/publications/book-orders. To submit
requests for commercial use and queries on rights and licensing, see https://www.who.int/copyright.

Third-party materials. If you wish to reuse material from this work that is attributed to a third party, such as tables, figures
or images, it is your responsibility to determine whether permission is needed for that reuse and to obtain permission from
the copyright holder. The risk of claims resulting from infringement of any third-party-owned component in the work rests
solely with the user.

General disclaimers. The designations employed and the presentation of the material in this publication do not imply
the expression of any opinion whatsoever on the part of WHO concerning the legal status of any country, territory, city or
area or of its authorities, or concerning the delimitation of its frontiers or boundaries. Dotted and dashed lines on maps
represent approximate border lines for which there may not yet be full agreement.

The mention of specific companies or of certain manufacturers’ products does not imply that they are endorsed or
recommended by WHO in preference to others of a similar nature that are not mentioned. Errors and omissions excepted,
the names of proprietary products are distinguished by initial capital letters.

All reasonable precautions have been taken by WHO to verify the information contained in this publication. However, the
published material is being distributed without warranty of any kind, either expressed or implied. The responsibility for the
interpretation and use of the material lies with the reader. In no event shall WHO be liable for damages arising from its use.

Design by Inis Communication

Contents

Acknowledgements iv

Abbreviations and acronyms vi

Glossary vii

1. Introduction 1
1.1 Background 1
1.2 Patient care pathway 2
1.3 Purpose 5
1.4 Target audience 5
1.5 Targets 5

2. Methodology 7
2.1 Stakeholder and Task Force consultation 7
2.2 Cost–effectiveness modelling 7
2.3 Delphi process and technical consultation 10
2.4 Public consultation and Scientific Target Product Profile Development Group meeting 11

3. Target product profiles 13

3.1 Use cases for tests for TB treatment monitoring and optimization 13
3.2. Minimal and optimal targets for key characteristics for tests for TB treatment monitoring
and optimization 15
3.3 Predictive values 21
3.4 Costs and cost effectiveness 23
3.5 Prioritization of test characteristics 25

References 27

Annexes 31
Annex 1. Declarations of interests 31
Annex 2. Results of the stakeholder consultation and Delphi survey 33
Annex 3. Technical consultation for the development of target product profiles
for tests and biomarkers for monitoring and optimizing tuberculosis treatment,
26–28 September 2022 (virtual meeting) 35
Annex 4. Scientific TPP Development Group meeting, 27–29 March 2023, Istanbul,
Türkiye (hybrid meeting with remote connection) 38

 iii
Acknowledgements

This document has been prepared by the Global Tuberculosis Programme of the World Health
Organization (WHO) with support from the Target Product Profiles (TPPs) Core Group, consisting
of Saskia den Boon (WHO, Switzerland), Claudia Denkinger (University of Heidelberg, Germany),
Dennis Falzon (WHO, Switzerland), Ankur Gupta-Wright (University College London, United Kingdom
of Great Britain and Northern Ireland, and University of Heidelberg, Germany) and Emily MacLean
(University of Sydney, Australia), with input from the TPP Task Force, consisting of Daniela Cirillo (San
Raffaele Institute, Italy), Frank Cobelens (Amsterdam Institute for Global Health and Development,
Netherlands (Kingdom of the)), Stephen Gillespie (University of St Andrews, United Kingdom),
Mikashmi Kohli (FIND, Switzerland), Morten Ruhwald (FIND, Switzerland) and Rada Savic (University
of California San Francisco, United States of America).

WHO thanks all other members of the Scientific TPP Development Group who met 27–29
March 2023: Mustapha Gidado (KNCV TB Plus, Netherlands (Kingdom of the)), Delia Goletti
(Translational Research Unit, National Institute for Infectious Diseases–Scientific Institute for Research,
Hospitalization and Healthcare [INMI-IRCCS], Italy), Rumina Hasan (Aga Khan University, Supranational
Reference Laboratory, Karachi, Pakistan, and London School of Hygiene and Tropical Medicine,
United Kingdom), Cathy Hewison (Médecins Sans Frontières [MSF], France), Kobto Koura (The
International Union Against Tuberculosis and Lung Disease [UNION], France), Christian Lienhardt (French
National Research Institute for Sustainable Development, France, and FAST-TB Initiative, CRDF Global,
USA), Patrick Lungu (East, Central and Southern Africa Health Community [ECSA], Zambia), Timothy
McHugh (University College London, United Kingdom), Lindsay McKenna (Treatment Action Group
[TAG], USA), Thomas Scriba (University of Cape Town, South Africa) and Christine Sekaggya-Wiltshire
(Infectious Diseases Institute, Uganda). Funding agencies were represented by the following: Grania
Brigden (The Global Fund to Fight AIDS, Tuberculosis and Malaria, Switzerland), Debra Hanna (Bill &
Melinda Gates Foundation, USA) and Cherise Scott (Unitaid, Switzerland).

WHO also thanks the modelling team that conducted a health economic analysis to inform the
development of the TPPs: Abdulkadir Civan (University of Heidelberg, Türkiye) and Florian Marx
(University of Heidelberg, Germany), with input from Hae-Young Kim (New York University, USA) and
Hojoon Sohn (Seoul National University, Republic of Korea).

WHO also appreciates the input of those who participated in the virtual technical consultation
held 26–28 September 2022: Macarthur Charles (Centers for Disease Control and Prevention [CDC],
USA), Keertan Dheda (University of Cape Town and London School of Hygiene and Tropical Medicine,
South Africa), Kathy Eisenach (independent consultant, USA), Ronald Allan Fabella (Disease Prevention
and Control Bureau, Department of Health, Philippines), Anneke Hesseling (Stellenbosch University,
South Africa), Ravinder Kumar (Central TB Division, National Tuberculosis Elimination Programme,
India), Yuhong Liu (Beijing Chest Hospital, China), Sanjay Kumar Mattoo (Central TB Division, National
Tuberculosis Elimination Programme, India), Norbert Ndjeka (National TB Programme, South Africa),
Ezio Tavora dos Santos Filho (WHO Civil Society Task Force and Rio de Janeiro Federal University,
Brazil), Boitumelo Semete-Makokotlela (South African Health Products Regulatory Authority,
South Africa), Jose Lapa e Silva (Ministry of Health, Brazil), Kelly Stinson (Cultura, LLC, USA), Nguyen
Thuy Thuong (Oxford University Clinical Research Unit, Viet Nam), Cesar Ugarte-Gil (Universidad
Peruana Cayetano Heredia, Peru) and Hui Xia (National Center for TB Control and Prevention, China
Center for Disease Control, China). The United States Agency for International Development [USAID],
USA was represented by Sevim Ahmedov.

iv Target product profiles for tests for tuberculosis treatment monitoring and optimization
WHO acknowledges the participation in this virtual technical consultation of several commercial
developers of tests for TB treatment monitoring and optimization: Devasena Gnanashanmugam
(Cepheid, USA), Ammar Jagirdar (former employee of Qure.ai, India), Nakaishi Kazunari
(Tauns Laboratories, Japan), Ahmed Maged (Abbott, USA), Megumi Komada (LSI Medience, Japan),
Jerome Nigou (Institut de Pharmacologie et Biologie Structurale, France), Akos Somoskovi (Roche,
USA) and Sruti Sridhar (Qure.ai, India).

WHO also acknowledges the participation of members of the WHO Secretariat: Nazir Ismail,
Fuad Mirzayev, Samuel Schumacher, Kerri Viney, Matteo Zignol (Global Tuberculosis Programme,
Switzerland), Martin van den Boom (Regional Office for the Eastern Mediterranean), Ernesto Montoro
(Regional Office for the Americas), Askar Yedilbayev (Regional Office for Europe), Kleydson Andrade
(Country Office for Brazil), Nkateko Mkhondo (Country Office for South Africa), Kirankumar Rade
(Country Office for India) and Chen Zhongdan (Country Office for China), as well as Corinne Merle
(Special Programme for Research and Training in Tropical Diseases). Overall guidance and direction
were provided by Tereza Kasaeva, Director of the Global Tuberculosis Programme.

The meeting, reviews and document were funded through a grant provided by USAID.

Acknowledgements v
Abbreviations and acronyms

COVID-19 coronavirus disease

DALY disability-adjusted life year
DR-TB drug-resistant tuberculosis
DS-TB drug-susceptible tuberculosis
mWRD molecular WHO-recommended rapid diagnostic test
NPV negative predictive value
NTPs national tuberculosis programmes
PPV positive predictive value
SSM sputum smear microscopy
TB tuberculosis
TPP target product profile
WHO World Health Organization

vi Target product profiles for tests for tuberculosis treatment monitoring and optimization
Glossary

Unless otherwise specified, the definitions listed below apply to the terms as used in this publication.
They may have different meanings in other contexts.
• Good outcome (also known as good treatment outcome) is considered to be bacteriological
or clinical improvement, or both, at the end of treatment for tuberculosis (TB) without evidence
of relapse within 6 months. This definition is aligned with revised World Health Organization
(WHO) treatment outcome definitions and incorporates the definition of “cured” (i.e. evidence of
bacteriological response) as well as includes people without bacteriologically confirmed TB who
have a good clinical response (1, 2). This definition also incorporates the operational research
definition of “sustained treatment success” (1).
• Poor outcome (also known as poor treatment outcome) is considered to be a lack of bacteriological
or clinical improvement, or both, by the end of TB treatment; early relapse; the need to prematurely
terminate or switch TB treatment; or death related to TB. This does not consider all post-treatment
complications or other aspects of cure that may be important to people with TB. It does not include
people who were lost to follow up from TB treatment, as it is unlikely any tests would be able to
predict who will be lost to follow up.
• Bacteriological response refers to bacteriological conversion of positive cultures (for drug-resistant
TB and drug-susceptible TB) or smears (for drug-susceptible TB only) to negative without reversion.
Reversion refers to cultures or smears becoming positive after bacteriological conversion.
Bacteriological response is relevant only to those people with bacteriologically confirmed TB who
have had serial samples analysed.
• Early relapse is reversion of bacteriological response or recurrence of TB symptoms in those who
have completed TB treatment or been declared cured within 6 months of the end of TB treatment.
• TB treatment optimization refers to initiating or switching to an effective TB treatment regimen
that results in a high likelihood of a good outcome. This includes using treatment stratification at
treatment initiation when it is determined that some people could achieve a good outcome on a
less intensive regimen (which may be a shorter regimen or a regimen with fewer medicines), while
others might need a more intensive regimen (which might be longer, include more medicines or
include adjuvant interventions) to achieve a good outcome. It also includes changing treatment as
a result of poor response to the current treatment regimen.
• Test refers to biomarker-based and non-biomarker-based tests, such as imaging-based tests, a score
based on clinical features or an assessment of cough sounds. While many of the characteristics
and targets in these TPPs assume that a test (or tests) will be biomarker based, the TPPs also apply
to non-biomarker-based tests. Such tests may even be more acceptable to people with TB and
health care workers and preferred to biomarker-based tests.

Glossary vii
1. Introduction

1.1 Background
Tuberculosis (TB) continues to be a major cause of morbidity and mortality globally despite being
curable and preventable. In 2021, an estimated 10.6 million people developed TB disease, and an
estimated 1.6 million people died from TB (3). Recent increases in the incidence of and mortality
from TB after many years of steady declines have been attributed to disruptions associated with the
coronavirus disease (COVID-19) pandemic. Although the largest gap in the TB care cascade remains
between the estimated incidence and number of cases notified – that is, an estimated 4.2 million
people with TB who are not notified and thus probably not diagnosed or treated – worldwide
only 86% of people who started on TB treatment in 2020 successfully completed it. Treatment success
remains lower in some World Health Organization (WHO) regions than in others (e.g. it is 72% in the
Region of the Americas), in people living with HIV (77% globally in 2020) and in those with drug-
resistant TB (DR-TB), for whom it is 60% (3).

TB treatment regimens are long and arduous for people with TB and can be associated with both
serious adverse effects and significant financial costs, which impact adherence and treatment
outcomes (4). Monitoring TB treatment to identify those who are at risk of poor outcomes could
improve overall treatment success. WHO currently recommends the regular use of sputum smear
microscopy and mycobacterial culture to monitor the response to treatment in adults with pulmonary
TB (4–6). However, these methods have important limitations as they rely on people on treatment
producing sputum samples and, therefore, are less useful in certain populations (e.g. in people
with extrapulmonary TB, children and people living with HIV). Sputum smear microscopy has poor
diagnostic accuracy for predicting poor outcomes, and culture is expensive, slow and not readily
available in many settings with a high burden of TB (7). While rapid, near-patient molecular assays
to detect Mycobacterium tuberculosis have transformed the diagnosis of TB, these assays are not
currently suitable for monitoring the response to TB treatment due to the persistence of nucleic acids
in nonviable TB bacilli during and sometimes beyond successful TB treatment (8).

WHO guidelines also recommend sputum smear microscopy or mycobacterial culture, or both, as
a test of cure, and the treatment outcome definition of cure includes a negative sputum culture or
smear during the last month of treatment (1, 4–6).1 Relying on sputum samples as tests of cure for
TB has limitations similar to those for monitoring TB treatment, with even fewer people being able
to produce quality samples at the end of treatment (9, 10). The current definition of cure for routine
programmatic monitoring also does not consider relapse, and sputum culture at the end of treatment
has poor sensitivity for relapse (11, 12). Although other indicators (e.g. weight, clinical symptoms,

1
Prior to 2021, cure was defined by WHO as a “pulmonary TB patient with bacteriologically confirmed TB at the beginning of treatment
who was smear- or culture-negative in the last month of treatment and on at least one previous occasion”. Since 2021, cure is now
defined as a patient who has “completed treatment as recommended by the national policy with evidence of bacteriological response
and no evidence of failure”. Bacteriological response is defined as the conversion of sputum culture or smear without reversion.

1. Introduction 1
chest X-ray or other imaging) are recommended in some guidelines to monitor TB treatment, including
in children and in extrapulmonary or sputum-negative TB disease, there is a need for accurate tests
to identify people who have been cured of TB (5, 6, 13, 14).

Clinical trials of novel, shorter TB treatment regimens for drug-susceptible TB (DS-TB) have consistently
shown that the vast majority of people with TB achieve relapse-free cure with 4 months of treatment
(15). This suggests that the standard 6-month first-line regimen is longer than needed for most people
with TB to achieve sustained cure (i.e. most people with TB are overtreated). A 4-month regimen
composed of rifapentine, isoniazid, pyrazinamide and moxifloxacin is now conditionally recommended
by WHO as an alternative to the current standard 6-month regimen (5, 16). A 4-month regimen has
also been recommended for non-severe TB in children and young adolescents, with non-severe being
defined by chest X-ray findings and clinical signs (16, 17). An alternative strategy of treating people
with TB with even shorter regimens (i.e. 2 months of treatment with bedaquiline and linezolid added
to isoniazid, pyrazinamide and ethambutol) and then treating those who relapse for longer was non-
inferior for a composite outcome of death, ongoing treatment or active disease when compared with
the standard 6-month treatment in a clinical trial (18). WHO target regimen profiles for TB treatment
aim for new TB regimens to be 2 months or shorter for both DS-TB and DR-TB (19).

Being able to accurately predict who will achieve a good outcome with shorter treatment regimens and
who will need longer or different regimens could have important clinical and public health benefits,
and could improve the quality of care for individual people with TB (20). This is even more pertinent
for DR-TB, for which a range of regimens with differing durations and adverse effects are currently
recommended (5). People with TB differ in their risk of relapsing after treatment. Therefore, there is
a clear need for new tools that can accurately predict outcomes in people who are starting or already
taking TB treatment and that allow for treatment to be optimized to improve outcomes.

There are novel platforms and biomarkers in the pipeline that offer the potential to monitor
TB treatment, predict outcomes, identify cure and allow optimization of management. These have
been summarized in reviews (21, 22) (a detailed review is beyond the scope of this document), and
potential tests include:
• host characteristic assays, including assays for cytokines, transcriptomic profiles and other biomarkers;
• pathogen burden and fitness assays;
• imaging-based assays;
• clinical scores, clinical symptoms and signs, cough sounds and lung function tests.

1.2 Patient care pathway

Integrating new technologies and diagnostics into care pathways is complex. Mapping and
understanding care pathways can inform the best fit for new technologies within a pathway and can
identify opportunities for improvement, cost savings and more efficient processes in existing pathways
(23). Care pathway analysis has been cited as a way of improving the downstream impact of possible
new tests on outcomes for people with TB (24). For these target product profiles (TPPs), TB treatment
guidelines from international agencies and countries with a high burden of TB were reviewed to
develop a diagrammatic visualization of current TB treatment care pathways and to describe the main
clinical decisions and questions associated with treating individual people with TB. A literature review
and survey of national TB programmes (NTPs) were undertaken to understand how well current TB
treatment monitoring tests and tests for cure are implemented, and which implementation barriers
need to be overcome for the next generation of tests to have a positive impact on patient care.

2 Target product profiles for tests for tuberculosis treatment monitoring and optimization
Fig. 1 presents a diagrammatic representation of a typical care pathway for people diagnosed
with DS-TB; it is based on a review of international TB treatment guidelines, including those from
countries with a high burden of TB, as well as the published literature on the implementation
of sputum-based tools for monitoring TB treatment.

The results of the review were supported by a survey about the implementation of monitoring for TB
treatment sent to NTPs and others who support TB treatment programmes in high-burden countries. The
results demonstrate that TB treatment monitoring is implemented in most settings using sputum smear
microscopy and culture, and, to a lesser extent, chest X-ray, clinical assessment and a patient’s weight.
Monitoring most commonly occurs at the primary care level, with visits at 2 months after treatment
initiation and at the end of treatment; note that monitoring for treatment of DR-TB usually occurs
monthly at the secondary care level. The main barriers to implementing treatment monitoring were
cost and laboratory capacity to support the tests, delays in receiving test results, difficulties in people
attending health care facilities for follow up and the workload for health care workers looking after
people with TB. Only a small minority of countries have started to implement 4-month TB treatment
regimens for DS-TB, and no settings have yet considered how monitoring will be adapted for newer,
shorter regimens. These factors were considered while developing the characteristics and targets for
these TPPs.

1. Introduction 3
Fig. 1. Graphical representation of a current typical care pathway for treating people
with drug-susceptible TB in high-burden countries
Current TB care pathway Notes New tests for treatment
monitoring and
optimizationa

Decision made to Based on positive sputum Tests for TB diagnosis could be

treat for TB tests (i.e. SSM, mWRD, integrated with tests for TB
culture), non-sputum tests, treatment optimization
chest X-ray or a clinical TB
diagnosis
DR-TB treatment and Yes Evidence of drug Tests for TB treatment
management resistance? Informed by rapid tests for optimization may need
drug resistance baseline measurements when
treatment is commenced
No
Decisions to be made by
Initiate standard health care worker about TB Tests to identify people who
treatment for DS-TB management; for children, should not receive a less
who can be started on intensive TB regimen
4-month regimen for could guide:
non-severe TB based on chest • DS-TB regimen;
X-ray and/or clinical signs • follow up, treatment support;
• In-person support?
• Follow up? • adjunctive therapies.
• 4-month regimen?

Baseline visit(s)
2-month visit

Sputum for SSM with Recommended at 2 months, Identify patients with a poor
or without culture although suboptimally response to treatment
implemented to inform the decision to step
down to continuation phase or
switch to different TB regimen
• Check adherence and as new regimens become
Decision informed by results of
adherence interventions available (e.g. from less to
Step down to SSM and culture, severity of
• Review diagnosis No more intensive regimen); can
continuation phase of disease (e.g. SSM grade, chest
• Refer to specialist be done days or weeks after
treatment? X-ray), improvements in
• Repeat tests for drug treatment initiation
symptoms, weight gain,
resistance
adherence and treatment
• Restart TB treatment
support, comorbidities
Yes

5 and/or 6-month visit

Recommended at 5 and 6 Tests for treatment

Sputum for SSM with
months in many guidelines, optimization at the end of TB
or without culture
although suboptimally treatment will inform the
Treatment failed implemented decision to complete or stop
• Restart TB treatment treatment. A test that indicates
• Check adherence and poor or inadequate response
adherence interventions No Complete treatment Decision informed by results of could lead to continuation of
• Review diagnosis at end of regimen? SSM at 2 and 5 or 6 months, treatment, a change in regimen
• Refer to specialist adherence, severity, response to or further testing
• Repeat tests for drug treatment, chest X-ray,
resistance Yes comorbidities

Successful treatment
completion Cured or treatment completed as
per WHO guidelines

Follow up and
management for
post-TB lung disease

DR-TB: drug-resistant TB; DS-TB: drug-susceptible TB; mWRD: molecular WHO-approved rapid diagnostic test; SSM:
sputum smear microscopy; TB: tuberculosis.
a
The timing for new tests for TB treatment monitoring and optimization are likely to change, depending on test
characteristics and new TB regimens. Monitoring tests for TB treatment may be done before the 2-month treatment visit.

4 Target product profiles for tests for tuberculosis treatment monitoring and optimization
1.3 Purpose
The overall purpose of these TPPs is to provide a set of parameters to guide the development and
manufacture of new tests to monitor and optimize TB treatment while considering the needs of
TB programmes and people with TB. While tests for TB optimization may well be useful in clinical
trials of TB treatment (25), the primary purpose of these TPPs is to develop tests for programmatic
use in high-burden settings. In parallel, Maclean et al. have developed guidance for generating
evidence to advise researchers about how to evaluate candidate tests (MacLean EL, et al., manuscript
in preparation, 2023). Recent advances in TB treatment regimens and important uncertainties about
the potential role for novel tests in monitoring and optimizing TB treatment have created a demand
for these TPPs. However, it is acknowledged that they may need reviewing and updating before the
typical 5-year period ends.

1.4 Target audience

The target audience for these TPPs includes commercial test developers and manufacturers,
people with TB, and members of academia and research institutions, regulatory agencies,
nongovernmental organizations, private sector implementers, NTPs, and civil society organizations,
and donors.

1.5 Targets
These TPPs provide both minimal and optimal targets for each included characteristic. The minimal
requirements are the lowest acceptable level for that characteristic, and the optimal requirements are
the ideal levels for that characteristic, expected to have the greatest public health impact (Table 1).
The minimal and optimal targets represent a range. Ideally, products will meet all minimal targets
and as many of the optimal targets as possible.

Table 1. Definitions of target product profile targets

Term Definition
Characteristic A test requirement or specification that is measurable
Minimal For a specific characteristic, minimal refers to the lowest acceptable output for that
characteristic. To be acceptable, tests should meet the minimal target.
Optimal For a specific characteristic, optimal represents the ideal output that is believed to be
realistically achievable. Meeting the optimal targets will have the greatest impact for
end-users, clinicians and people with TB. Ideally, developers would design and develop
their solutions to meet the optimal targets for all characteristics.

1. Introduction 5
2. Methodology

WHO followed a stepwise approach to identify test characteristics that are important for people with
TB and TB programmes, as well as for test developers and other stakeholders.

A TPP Core Group was formed to coordinate the development of the TPPs and lead the writing process.
WHO also constituted a Scientific TPP Development Group, consisting of leading scientists and experts,
public health officials, and in-country end-user representatives. Members of this group were engaged
throughout the TPP development process and proposed the final TPPs during an in-person meeting.
Members of the Scientific TPP Development Group completed standard WHO declarations of interest
procedures (Annex 1). A sub-set of this group formed the TPP Task Force and was consulted more
frequently to help direct the TPP development.

The initial draft TPP document was prepared by the TPP Core Group based on a systematic literature
review (22), a draft (unpublished) TPP developed by FIND, and multiple meetings of the Core Group
and the TPP Task Force.

The process further included an online stakeholder consultation, cost-effectiveness modelling a

modified Delphi process, a technical consultation,and an online public consultation, as described below.
The final TPPs were agreed on during a meeting of the Scientific TPP Development Group.

2.1 Stakeholder and Task Force consultation

An online stakeholder consultation was held to canvass the opinions of a wide range of stakeholders,
including test developers, researchers, staff of NTPs, laboratory scientists, implementers and clinicians,
and representatives from industry and civil society organizations. Stakeholders were asked about the
perceived need for the TPPs; categories, definitions and aims of the tests; target patient populations;
and key characteristics of the tests. Additional comments were sought when there was disagreement
with the proposed definitions or characteristics. A brief summary of the findings of the stakeholder
consultation is provided in Annex 2. The draft TPPs were revised based on the stakeholder consultation
and changes suggested during Task Force meetings (version 0.0, September 2022).

2.2 Cost–effectiveness modelling

Integrating early economic evaluation into TPPs is useful to guide the development of target criteria
and to inform product research and development.

A Markov multistate model was developed to explore the potential health impacts and costs of
hypothetical novel tests used during TB treatment to identify people at high risk of a poor treatment
outcome2 who would benefit from further investigation or extended or modified treatment, or a

2
Tests used for treatment monitoring are assumed to identify patients for whom treatment produces an inadequate response and who
are therefore at high risk of a poor treatment outcome.

2. Methodology 7
combination of these (i.e. treatment monitoring; see also Section 3.1 about use cases). The model
was used to simulate standard treatment for non-rifampicin-resistant TB in a hypothetical cohort of
1 000 adults diagnosed with pulmonary TB.

The core model structure (Fig. 2) consists of three states used to distinguish between people who
will have an adequate response to treatment and those who will have an inadequate response due
to drug resistance, or poor adherence or other causes.3 People initiating TB treatment move into one
of the three states conditional on undetected drug resistance and other reasons. During the course
of treatment, people transition from an adequate response to an inadequate response to treatment
at rates describing the incidence of poor adherence and the acquisition of DR-TB. An inadequate
response due to poor adherence includes individuals who take their medication irregularly and those
who interrupt their treatment.

The model’s structure includes subdivisions for time spent with an adequate response to treatment (0
to 6 months; Fig. 2). People who complete the final month of treatment transition into one of two
outcome states, cure or failure/relapse, at probabilities depending on the total time spent with an
adequate response to treatment. People with fewer months of an adequate response have a lower
probability of cure compared with those with up to 6 months of adequate response. Failure/relapse
is a composite outcome that includes people with persistent active disease during treatment or early
disease reactivation (i.e. relapse). Cure is assumed in the absence of failure and of relapse during a
period of at least 2 years after treatment completion. Individuals can die at any time during treatment,
with higher rates of death assumed for people during their first month of treatment and among those
with an inadequate response to treatment.

Treatment monitoring is implemented as a process that can be enabled (i.e. switched on or off) at the
end of each month of treatment (Fig. 2). A probabilistic decision-analytic process is implemented to
simulate interventions after a positive test. These consist of (i) adherence assessment and counselling
plus a 2-month extension of treatment (i.e. in the event of poor adherence) and (ii) drug-resistance
testing and initiation of treatment for DR-TB (i.e. if resistance is confirmed).

Four model scenarios were considered for the analysis (Table 2). The base-case scenario assumes no
treatment monitoring. Each test scenario assumes a reference test is conducted prior to initiation
of treatment.

Monitoring scenario 1 assumes the use of smear microscopy to detect an inadequate response
to treatment. Two additional scenarios assume the use of hypothetical novel tests that meet the
minimal (scenario 2) and optimal (scenario 3) criteria for sensitivity and specificity to detect an
inadequate response to treatment.

Costs incurred under each scenario are estimated in 2023 US dollars (US$), adopting a health care
system perspective. Average costs reflect the costs of conducting a particular monitoring test,
assessing adherence, providing counselling, conducting drug-resistance testing (i.e. in the event
of a positive monitoring test) and providing treatment. Model parameters were obtained from the
published literature. The number of unfavourable treatment outcomes averted (i.e. death, failure/
relapse) and disability-adjusted life years (DALYs) averted were estimated to measure the health impact
achieved under the monitoring scenarios. Costs and DALYs averted were discounted at an annual
rate of 3.0%. Incremental cost–effectiveness ratios were estimated using fixed estimates of costs per
test performed. In addition, for each hypothetical test scenario, costs per monitoring test performed
were estimated in relation to variable willingness-to-pay thresholds.

3
Other causes of an inadequate response to treatment may include, for example, malabsorption of the medicine or severe TB disease
resulting in a delayed treatment response.

8 Target product profiles for tests for tuberculosis treatment monitoring and optimization
The model was used to explore the potential health impacts and to estimate incremental cost–
effectiveness ratios of hypothetical monitoring tests (Table 2). Incremental cost–effectiveness ratios
represent the additional costs incurred per additional DALY averted, using smear microscopy as
a reference. These ratios were estimated for different price levels per monitoring test performed and
compared against different willingness-to-pay thresholds, including the estimated cost per DALY
averted with monitoring using smear microscopy. The model uses probabilistic analysis to account
for parameter uncertainty. Best estimates represent the mean of 1 000 model trajectories with 95%
uncertainty intervals calculated as the 2.5th and 97.5th percentile values.

Fig. 2. Structure of the Markov multistate model used to assess the potential health
and cost–effectiveness impacts of hypothetical novel tests used during TB treatment to
identify people at high risk of a poor outcome

Core TB treatment model

Adequate response to
treatment

Treatment for
drug-resistant TB
Monitoring t

Inadequate response to Inadequate response to

treatment due to poor treatment due to
adherence and other causesa drug resistance

Failure/
Cure
relapse

Subdivisions: months of TB treatment completed with adequate response

0 1 2 3 4 5 6

TB: tuberculosis; t: time of treatment monitoring.

a
Other causes of an inadequate response to treatment may include, for example,
malabsorption of medicines or slow treatment response due to severe TB disease.

2. Methodology 9
Table 2. Scenarios and their key parameters used in the Markov multistate model
to assess the potential health and cost–effectiveness impacts of hypothetical novel
tests used during TB treatment to identify people at high risk of a poor outcome
Description Mean Uncertainty Source
value interval
Base-case scenario
(no monitoring)
Monitoring scenario 1
(Sputum smear microscopy, month 2)
Monitoring coverage (sputum-based) 0.750 0.650–0.850 Assumptiona
Sensitivity 0.448 0.225–0.673 Model estimate, based on
Horne et al. 2010 (26), Davis et al. 2013 (27)
Specificity 0.826 0.732–0.924 Model estimate, based on
Horne et al. 2010 (26), Davis et al. 2013 (27)
Cost per testb 5.00 2.00–8.00 Unit cost study repository
(https://www.ghcosting.org)
Monitoring scenario 2c
(Hypothetical test with TPP minimum criteria, month 2)
Monitoring coverage (sputum-based) 0.750 0.650–0.850 Assumptiona
Sensitivity 0.825 0.750–0.900 TPP test criteria (minimal)
Specificity 0.850 0.800–0.900 TPP test criteria (minimal)
Cost per testb Variesd NA NA
Monitoring scenario 3 c

(Hypothetical test with optimum criteria, month 2)

Monitoring coverage (sputum-based) 0.750 0.650–0.850 Assumptiona
Sensitivity 0.925 0.900–0.950 TPP test criteria (optimal)
Specificity 0.925 0.900–0.950 TPP test criteria (optimal)
Cost per test b
Varies d
NA NA

NA: not applicable; TPP: target product profile.

a
Sputum-based monitoring tests have lower test coverage due to the inability of some people with TB to produce sputum;
coverage of 65–85% is assumed, reflecting assumptions about the frequency of sputum sparsity.
b
Costs per test are in US dollars.
c
Monitoring scenarios 2 and 3 are subdivided into scenarios of sputum-based tests and non-sputum based tests; non-
sputum-based tests are assumed to have a higher monitoring coverage, in the range of 85–95%.
d
Costs for hypothetical monitoring tests are varied in the analysis to identify ranges that meet thresholds for cost effectiveness.

2.3 Delphi process and technical consultation

Version 0.0 of the TPP document was reviewed during a technical consultation with a wide range
of stakeholders, consisting of product developers, scientists, researchers, NTP managers, implementers
and representatives from civil society organizations (Annex 3). Ahead of the meeting, participants were
sent the draft v. 0.0 TPP and a Delphi survey. Participants were asked to reply to specific questions,
expressing their level of agreement with the proposed targets according to a predefined Likert scale
ranging from 1 to 5 (1 – agree, 2 – somewhat agree, 3 – neither agree nor disagree, 4 – somewhat
disagree and 5 – disagree). Individuals were asked to provide comments or alternative targets when they
did not agree with a proposed target (i.e. those scored at 3, 4 or 5). The results (Annex 2) were used
to focus discussions at the virtual technical consultation. Input and comments from product developers
were considered in light of their affiliation. Changes and suggestions from the Delphi process and
technical consultation were incorporated into an advanced draft document (v. 0.1, November 2022).

10 Target product profiles for tests for tuberculosis treatment monitoring and optimization
2.4 Public consultation and Scientific Target Product Profile
Development Group meeting
Public comment on draft v. 0.1 of the TPPs was invited through an online survey that was distributed
through the WHO mailing list. The draft TPP and online survey were available from 9 February to 9
March 2023. The results of the public consultation were presented at a final Scientific TPP Development
Group meeting (Annex 4), at which agreement on the final TPPs was obtained (v. 1.0, May 2023).

2. Methodology 11
3. Target product profiles

3.1 Use cases for tests for TB treatment monitoring and optimization
Based on the decisions facing clinicians who manage people with TB, as outlined in the patient care
pathway in Fig. 1, and possible stratified TB treatment regimens in the future, three use cases for
tests for TB treatment monitoring and optimization have been identified. These use cases are for
tests conducted (i) at the time of TB treatment initiation, to identify people who will not achieve
a good treatment outcome (see Glossary) with a less intensive TB treatment regimen; (ii) during
TB treatment, to identify people at high risk of a poor treatment outcome and who would benefit
from further investigation or a more intensive TB treatment regimen, or both; and (iii) at the end of
TB treatment, to identify those who have not achieved a good treatment outcome. These use cases
are further described in Table 3.

Less intensive TB treatment regimens may be shorter in duration or contain fewer medicines. A more
intensive TB treatment regimen may be longer, contain more medicines or include adjuvant therapies. It
is assumed that these different regimens will be non-inferior to the standard of care in terms of efficacy,
but less intensive regimens will have some advantages to people with TB and TB programmes (e.g.
shorter duration, smaller pill burden, less risk of adverse reactions to medicines).

Table 3. Summary of use cases for tests for TB treatment monitoring and optimization
Use case Timing Explanatory notes Consequences
Identify people Treatment Prioritizes avoiding undertreatment If the test predicts a poor treatment
who require a initiation of people who would have a poor outcome, the person will be started on
more intensive TB treatment outcome on a less intensive a more intensive TB regimen, which may
treatment regimen regimen (e.g. those with more severe be longer or contain more medicines,
disease). The test would, therefore, aim or the patient may need additional
to predict with high accuracy those interventions.
likely to have a poor treatment outcome
If the test predicts a good treatment
on a less intensive treatment regimen
outcome, the person can be initiated on
(i.e. the tests need high sensitivity for
a less intensive TB regimen (albeit with
predicting poor outcomes).
ongoing monitoring).
Identify people During Aims to identify people who are not If the test shows a poor treatment
at risk of a treatment adequately responding to TB treatment. response or risk of poor treatment
poor outcome These are sometimes known as tests for outcome, the person may need a
with current TB TB treatment monitoring or treatment different, more intensive optimized
treatment response, and they have aims similar treatment regimen (e.g. longer or with
to using sputum-based microscopy or more medicines or adjuvant therapies),
culture during treatment. These tests may need to undergo further testing
need to be accurate enough to not (e.g. for drug resistance) or may need
miss people responding poorly and adherence support interventions or
also to minimize the number of people adjuvant therapies. If a test shows
incorrectly identified as at risk of a poor a good treatment response, the
treatment outcome. person can continue with the current
treatment.

3. Target product profiles 13

 Use case Timing Explanatory notes Consequences
Identify people with Presumed Aims to identify those who have a poor If the test shows a poor treatment
a poor treatment end of treatment outcome or are at risk of outcome or high risk of early relapse,
outcome at the end treatment early relapse (i.e. as is currently done the person may require further
of TB treatment using microscopy or culture). These tests investigations, continuing or optimizing
need to be accurate enough to not miss of the treatment regimen, or a
people who are not cured or who will combination of these.
relapse early and also to minimize the
number of people incorrectly identified
If the test shows a good treatment
as not cured. These are sometimes
outcome has been achieved, the current
called tests of cure.
regimen can be completed.

TB: tuberculosis.

• While there are many different aspects of TB disease severity and response to treatment (e.g.
bacteriological, clinical, radiological, functional), these tests will aim to identify markers that predict
the outcome of TB treatment and instances in which this outcome could be improved by modifying
treatment (i.e. in terms of medicines or length of treatment) or by adjuvant interventions.
• Ideally, one test will be developed that can be used for all use cases, but this is not mandatory.
The TPPs outline the characteristics that tests for all use cases will have in common, and they list
separately those characteristics for which the targets are different between use cases.
• Based on the TB care pathways for the 6-month DS-TB regimens currently implemented by TB
programmes in high-burden settings, follow-up visits and tests for TB treatment monitoring are
usually done at around 1–2 months after treatment initiation and tests of cure at around 5–6 months.
New tests may also be done several times and at different times, according to what they measure
and how they predict poor outcomes, and based on implementation of new treatment regimens
(e.g. the 4-month rifapentine-based regimen). TB programmes may change their TB care pathways
in response to newer regimens and diagnostic tests.
• In the patient care pathway analysis and cost–effectiveness modelling, the example of a 6-
month DS-TB scenario was used, but the TPPs also apply to the monitoring and optimization of
shorter regimens or those for DR-TB.
• Tests that identify drug resistance will also predict a poor outcome with standard DS-TB regimens,
and switching to appropriate TB regimens will improve outcomes. However, these TPPs are not
referring to tests of drug resistance. TPPs for tests of drug resistance are described elsewhere (28).
• A reference standard is useful when considering the diagnostic accuracy of tests. However, there is
no perfect test of treatment response or optimization. Therefore, the best proxy for the reference
standard for these tests will be the final treatment outcome 6 months after the end of treatment.
A more detailed discussion of reference standards will be available in a manuscript that provides
methodological guidance for evaluating tests for treatment monitoring and optimization (MacLean
EL et al., manuscript in preparation, 2023).

14 Target product profiles for tests for tuberculosis treatment monitoring and optimization
3.2. Minimal and optimal targets for key characteristics for tests for TB
treatment monitoring and optimization
Tables 4–6 describe the target characteristics for tests to be used when people start treatment, during
treatment and at the end of treatment; Table 7 describes the operational characteristics that all TPP
tests have in common.

Table 4. Target characteristics for tests used at treatment initiation to identify

people with TB who require a more intensive treatment regimen
Characteristic Minimal target Optimal target Explanatory notes
Sensitivity (for ≥90% ≥95% The test should be good at identifying people who are at high
detecting those risk for a poor treatment outcome if treated with less intensive
who require a regimens and who would have a better treatment outcome
more intensive with optimized TB treatment.
treatment
A less intensive TB treatment regimen may be shorter or
regimen)a
contain fewer medicines. The optimized TB treatment may be
longer, contain more medicines or include adjuvant therapies.
It is assumed that less intensive and optimized treatment
regimens will be non-inferior to the current standard of care
in terms of efficacy and that less intensive regimens will be
advantageous to people with TB and TB programmes (e.g.
because they are of shorter duration, use fewer medicines, and
have less toxicity and risk of adverse reactions).
This is a test to rule out those who will have a poor treatment
outcome on less intensive treatment and, therefore, it
prioritizes high sensitivity. Those identified as being at high
risk for a poor treatment outcome will not be initiated
on less intensive treatment regimens. Based on studies of
predictive models and scores, the sensitivities proposed are
reasonable but ambitious (29, 30). However, there may also be
opportunities later during treatment to identify those at risk of
a poor treatment outcome on less intensive regimens (i.e. early
during treatment monitoring). Note that some people will have
a poor treatment outcome due to poor adherence or other
factors that may not be predictable at treatment initiation.
Specificity (for ≥70% ≥80% The test should correctly identify those who could achieve a
detecting those good treatment outcome with less intensive treatment and
who can be thus help avoid unnecessary overtreatment.
treated with a
However, for most people overtreatment is considered less
less intensive
problematic than undertreating people for whom treatment
regimen)
would fail, or who would relapse or develop new resistance
with a less intensive regimen; therefore, for this use case,
sensitivity is prioritized over specificity.
Timing Prior to TB Up to 7 days Tests to identify people who are unsuitable for less intensive
treatment after starting TB regimens should be performed and provide results prior to
initiation treatment the patient starting treatment or at treatment initiation so that
treatment management can be optimized from the outset,
without adding extra follow-up visits.
However, the ideal test would still be accurate soon after TB
treatment initiation in case people are started on treatment
prior to a sample being collected for this test (e.g. if treatment
is initiated before a patient is referred to TB services where this
test is available).

TB: tuberculosis.
a
If a test detects those with a high likelihood of a good treatment outcome, the targets for sensitivity and specificity should
be reversed.

3. Target product profiles 15

Table 5. Target characteristics for tests used to identify people at risk of a poor
treatment outcome during TB treatment
Characteristic Minimal target Optimal target Explanatory notes
Sensitivity (for ≥75% ≥90% The test should be good at identifying people with inadequate
poor treatment response to treatment who are likely to have a poor treatment
response)a outcome on their current regimen. These people may benefit
from further investigation, optimization of TB treatment or
other interventions.
This test is important to improve programmatic treatment
success rates. The minimal target is based on the estimated
accuracy of the clinical models used to predict poor outcomes
in treating pulmonary TB as well as the accuracy of current
tools used for treatment monitoring.
Specificity (for ≥80% ≥90% The test should correctly identify those persons who are
good treatment responding to treatment and are likely to have a good
response) outcome on less intensive treatment so that their treatment
is not changed to a more intensive regimen unnecessarily.
This requirement is even more important if more intensive TB
treatment regimens have more adverse events or costs for
people with TB or health systems.
Timing For use halfway ≤ 4 weeks The results should be available as early as possible during TB
through the after starting treatment, but at a time when the test has sufficient likelihood
anticipated treatment of detecting poor treatment outcomes that could be prevented
TB treatment by optimizing TB treatment.
regimen
Programmatically, it would be advantageous if time points
corresponded to routine follow-up appointments. However,
the timing of routine follow up may change with newer and
shorter treatment regimens (e.g. the 4-month regimens already
recommended for adults and children) or improved diagnostics
for monitoring the response to and optimizing TB treatment,
or both.
The timing of the test might also depend on the type of test
(e.g. tests targeting mycobacteria may need to be done earlier
than tests targeting the host’s immune response). Timing may
also impact diagnostic accuracy as the risk of an undetected
poor outcome will be reduced over time. If additional
patient visits are required, this should be considered in cost–
effectiveness analyses.
Frequency For use at two For use once Ideally, the test would be done at only one time point to
time points (so at follow-up reduce costs and health service resources.
two results can visit during
It is likely the test will need to be done at least twice to
be compared) treatment
compare serial measurements (e.g. at treatment initiation or
(without
first follow up as a baseline measurement and then again at
the need
subsequent follow up). Ideally, the test would be the same test
for baseline
as the one used to identify people for less intensive regimens;
measurement)
if this is the case, the test done at treatment initiation could act
as the baseline value, and so the change between values at the
two time points could be considered.

TB: tuberculosis.
a
If a test detects those likely to have a good treatment outcome, the targets for sensitivity and specificity should be
reversed.

16 Target product profiles for tests for tuberculosis treatment monitoring and optimization
Table 6. Target characteristics for tests used to identify people with a poor treatment
outcome at the end of TB treatment
Characteristic Minimal target Optimal target Explanatory notes
Sensitivity ≥80% ≥95% The test should have high sensitivity to identify people who
(for detecting have a poor outcome despite completing their anticipated TB
those with a treatment. People identified as having a poor outcome will
poor treatment have further investigations or optimized treatment, or both,
outcome)a and be followed up appropriately.

Specificity (for ≥90% ≥95% The test should correctly identify those persons with a good
detecting those treatment outcome to reduce overtreatment, as this has
with a good important costs and consequences for people with TB and TB
treatment programmes. The prevalence of poor outcomes will be lower at
outcome) the end of TB treatment, which is why the specificity is higher
than for the other use cases.
Timing For use at the For use at the The test should be done at the time of or just before the
last possible last follow- anticipated completion of TB treatment to maximize its
point prior up visit or last accuracy for detecting poor treatment outcomes.
to the end of medication refill
Ideally, the test would be performed and still be accurate at the
TB treatment prior to the
final follow-up visit or final medication refill because, from an
(e.g. within the anticipated end
operational perspective, people may not return at the end of
last week of of TB treatment
treatment.
treatment)
Ideally, the timing should reduce the risk of people with a
poor outcome having a break in their treatment while waiting
for results; therefore, the timing of the test also requires
consideration of its turnaround time. Note that the duration of
treatment may vary as newer TB regimens are introduced or for
DR-TB or central nervous system disease.
Frequency For use at two For use only The test needs to identify people with a poor treatment
time points (so once, at outcome at the end of TB treatment.
two results can the end of
Ideally, the test would be done only once, at the end of
be compared) treatment
treatment, to reduce costs and resources. It is possible the
(without
test will need to be done at least twice to compare serial
the need
measurements (e.g. at treatment initiation and early during
for baseline
treatment). Ideally, the test would be the same test as for one
measurement)
or more of the other use cases; therefore, at least one other
result would be available for comparison.

DR-TB: drug-resistant TB; TB: tuberculosis.

a
If a test detects those with a good treatment outcome, the targets for sensitivity and specificity should be reversed.

3. Target product profiles 17

Table 7. Operational characteristics common to all tests described in these target
product profiles
Characteristic Minimal target Optimal target Explanatory notes
Assay or For use as an For use as a The tests should ideally be instrument-free, feasible for use
instrument instrument- point-of-care at the point of care and easily deployed at the primary care
design based test test that can be level, where most people with TB are seen, and tests should
requiring basic conducted at not require laboratory facilities or technicians. However,
laboratory primary care or instrument-based tests requiring basic laboratory infrastructure
infrastructure community level (e.g. as found at a district-level hospital) may be acceptable
(e.g. in a as a minimal requirement, as long as this does not make
laboratory at turnaround times too long.
a district-level
If the test requires the use of an instrument, this would
hospital or in
ideally be compact and have appropriate features to allow for
an equivalent
placement in peripheral or community health care facilities (e.g.
setting where
a small portable, battery-powered instrument). Tests requiring
molecular
instruments would ideally be suitable for multiple uses (e.g. for
WHO-
TB diagnostic testing and testing for other diseases) and should
recommended
be stand-alone devices not needing separate computers.
rapid diagnostic
tests can be Tests conducted by health care workers (rather than laboratory
done) staff) should also have a minimal impact on the health care
facility or setting where they are being implemented (i.e. they
should be fast, simple to operate and possibly come with an
automated reader).
Non-laboratory-based tests (e.g. imaging or clinical scores) can
also meet the requirements of the TPPs.
Target For use in a For use in a While TB diagnostics are sometimes centralized to higher-level
placement for district hospital– peripheral-level facilities, these tests should ideally be feasible to implement
test level laboratory health facility at all levels, including community health care facilities where
(i.e. without a people usually have their treatment and follow up.
laboratory)
The test would optimally be a point-of-care test that could be
implemented anywhere. Some tests may need a laboratory;
therefore, implementation at the district-laboratory level may
be acceptable, given that people with TB are already in care
and could potentially wait until their next visit for results,
although this is not ideal.
Target user of For use by a Health care Most TB treatment is managed and monitored in peripheral
test health care worker health care settings by health care workers with minimal
worker (or or patient training, including community health workers. Therefore, these
someone with (self-test) tests should ideally be targeted at the same level of health care
equivalent personnel.
training in
For TB treatment optimization, a laboratory-based test would
using molecular
also be acceptable, with samples being transported to the
WHO-
nearest laboratory.
recommended
rapid diagnostic The optimal target includes a self-test, whereby the people
tests) with TB performs the test and either interprets the results or
contacts a health care worker for interpretation or follow-up
action.

18 Target product profiles for tests for tuberculosis treatment monitoring and optimization
Characteristic Minimal target Optimal target Explanatory notes
Target For use by For use by all These tests are aimed at people starting or already receiving
population people with people starting TB treatment and at health care workers seeking to optimize
pulmonary TB or already subsequent treatment.
or any form of receiving
Ideally, the test would be applicable to all people starting
bacteriologically treatment for
TB treatment, including children, elderly people, pregnant
confirmed TB TB
women, people living with HIV, people with severe malnutrition
or comorbidities, and people with DR-TB or extrapulmonary
or disseminated TB, or both. However, it may be that a
test is applicable only to a subpopulation of people being
treated for TB (e.g. those with pulmonary TB or those with
bacteriologically confirmed TB), as current monitoring tests are.
Although not ideal, it may be acceptable if the nature of the
test precludes its use in all people, but it provides substantial
improvement over existing similar tests (e.g. sputum smear
microscopy) and its use is still cost effective; however, this may
impair uptake and implementation of the test.
Sample type (if For use with For use with Optimally, tests should use not only sputum as the sample type,
a clinical sample a minimally a minimally as sputum is not produced by all people with TB (particularly
is required) invasive invasive, easily subgroups such as children, people living with HIV and those
respiratory accessible with extrapulmonary TB), it is difficult to obtain safely and it
sample (i.e. sample (e.g. tends to become scarcer as TB treatment progresses.
not limited to urine, breath,
As a minimal requirement, tests should use respiratory samples
sputum alone) capillary blood)
and not be limited to sputum. Tests based on respiratory samples
(e.g. oral swabs, saliva or breath) are likely to have a higher yield
and thus perform better for people with pulmonary TB.
Ideally, samples that are minimally invasive and easy to access
are preferred, such as urine or capillary blood. The volume
required should be reasonable to collect with one sample.
Samples should require minimal processing or preparation prior
to testing, and they should pose minimal risks to health care
workers with respect to infection prevention and control.
Tests may not require clinical samples at all (e.g. imaging
modalities such as chest X-ray or digital chest X-ray with
computer-aided detection, ultrasound) or may be scores based
on multiple clinical observations, with or without extra clinical
samples being required. Tests with artificial intelligence–based
biomarkers of voice or cough sounds also would not require
clinical samples.
Time to result ≤1 day ≤2 hours The time to result reflects the time from when a sample is
received to the release of results under optimal programmatic
conditions.
Most people with TB will have treatment initiation and follow
up at primary care–level health facilities. Ideally, the results
should be available during the same clinical encounter so
health care workers can make decisions about management
and treatment immediately; this is particularly important for
tests done at the time of treatment initiation.
If it takes longer to receive the results, then people may need
to be contacted (e.g. by phone or SMS) for a return visit,
implying more cost for people treated for TB and risk of loss to
follow up. This may be necessary for tests requiring samples to
be transported to laboratories. Complementary measures will
need to be in place to facilitate sample transportation to the
place of testing and to automate the production and electronic
transmission of results to the clinician and patient.

3. Target product profiles 19

Characteristic Minimal target Optimal target Explanatory notes
Results Requires skills Easily Results should be easy to interpret by the target users for the test.
interpretation to interpret interpretable
When test results include monitoring a change in test results
by health care
or a parameter over time, there needs to be guidance for the
workers and
health care worker about how to interpret the results and
people with TB
translate these into changes in TB treatment.
Results should be easily available to the health care worker
and digitized so they can be transmitted to people treated for
TB (e.g. via SMS or mobile phone, or by e-mail). Whenever
possible, digitization of test results should be automated (e.g.
machine-to-machine or machine-to-server) to prevent the
need for data entry or transcription. Lateral flow devices could
come with readers for this purpose. The application of digital
technologies and artificial intelligence may be considered to
minimize the need for humans to interpret results (e.g. through
automated reading based on a digital photograph of lateral
flow tests or line-probe assay strips with machine learning
used to identify the presence or absence of bands, leading to
algorithmic interpretation of results). Instruments or assays that
produce quantitative data should come with clear guides to aid
interpretation by health care workers (e.g. a cut-off defining a
poor response to treatment).
Sample ≥1 test run ≥4 tests run If instruments are required, they should have sufficient capacity
throughput (for simultaneously independently to run multiple tests at the same time. They should also be
instrument- from one capable of running tests independently from one another so
based assays) another that turnaround time is not delayed.
The optimal target of ≥4 tests is based on the capacity of
platforms such as for the GeneXpert system for TB diagnosis
(Cepheid, Sunnyvale, CA, USA).
Operational Up to 30 °C Up to 45 °C Because countries with a high burden of TB commonly
environment and 70% and 90% experience high temperatures and high humidity, tests
relative relative and instruments should ideally be able to operate in such
humidity humidity, environments without requiring air conditioning.
no power
Instruments should also be robust enough to operate in dusty
requirements
environments if placed in clinics. Ideally, instruments and
devices should have a rechargeable battery to compensate for
an unreliable electricity supply without impacting operation
(e.g. a battery that needs to be charged only once every
24 hours).
Environmental Minimize adverse impacts on the Tests and any associated instruments should have minimal
impact environment adverse impacts on the environment.
This consideration includes the potential to produce tests
locally, to minimize waste and maximize reusability and the
recycling of by-products, to ensure that platforms have multiple
uses, the ability to recycle instruments at the end of their life,
to ensure low power consumption and radiation emissions, and
to use technology that relies on solar panels. Manufacturers
should be responsible for take-back, recycling and disposal
(e.g. by covering the associated costs). Manufacturers should
provide clear specifications for biomedical waste management
and infection control guidelines for during and after use and
for safe disposal.

20 Target product profiles for tests for tuberculosis treatment monitoring and optimization
 Characteristic Minimal target Optimal target Explanatory notes
Maintenance Yearly servicing Ideally, The need for instrument maintenance should be minimal,
(for instrument- of instrument maintenance requiring at most yearly maintenance and minimal expertise for
based assays) free, or maintenance and service.
maintenance
Maintenance and technical support should be available
should be
in-country or local staff could be trained to provide this, or
done locally or
both. Alternatively, maintenance could be done remotely. It
remotely
should be possible to receive software updates for instruments
over a low-bandwidth mobile internet connection. The cost of
maintenance should be low, and ideally, service agreements
will be detailed in the purchase contract.
Quality control Provision of Integrated Quality control would ideally be integrated into the test,
reagents for quality control including point-of-care tests.
quality control
If external reagents are required for quality control, these
should be provided with the test kits.
Training and ≤3 days ≤1 day Ideally, training for those performing the test should be minimal.
education
Optimally, health care workers at community or peripheral
health facilities should be able to conduct the tests with brief
training, but tests requiring training as a laboratory technician
would also be acceptable.

DR-TB: drug-resistant TB; TB: tuberculosis; TPP: target product profile.

3.3 Predictive values

Tables 4–6 include the targets for diagnostic accuracy in terms of sensitivity and specificity for
tests addressing the three key use cases of interest. However, from the patient’s, clinician’s and
programmatic perspectives, the predictive values of tests are more useful because they help understand
how likely or unlikely the disease state is, given a particular test result. Predictive values vary according
to the prevalence of poor treatment outcomes. Predictive values for each of the three use cases that
support the diagnostic accuracy targets are discussed below.

3.3.1 Tests to identify people at initiation of TB treatment who require a more

intensive regimen
Assuming a prevalence of poor treatment outcomes (i.e. either treatment failure or early relapse)
of 15% (1) and a test meeting the minimal target diagnostic accuracy of 90% sensitivity and 70%
specificity, the negative predictive value (NPV) will be 97.5% and the positive predictive value (PPV)
will be 34.6%. The impacts of a higher or lower prevalence are shown in Fig. 3. The high NPV is
important to reduce the number of people with TB being undertreated.

In a hypothetical cohort of 1 000 people with TB and using a test meeting the minimal TPP targets
for diagnostic accuracy, of the 150 who will have a poor treatment outcomes, 135 would be correctly
identified by this test and 15 would be missed (i.e. inappropriately started on a less intensive regimen).
Of the 850 who would have a good treatment outcome on a less intensive regimen, 595 would be
correctly identified and started on a less intensive regimen, while 255 would be incorrectly classified
as likely to have a poor treatment outcome and started on a more intensive regimen. Under the
optimal TPP diagnostic accuracy targets (≥95% sensitivity and ≥80% specificity), only 7 people likely
to have a poor treatment outcome on less intensive TB treatment would be missed, and 170 likely
to have a good treatment outcome on less intensive TB treatment would be overtreated with a more
intensive regimen.

3. Target product profiles 21

Fig. 3. Negative predictive values (NPVs) and positive predictive values (PPVs) for a
test with 90% sensitivity and 70% specificity for detecting people at risk of a poor
TB treatment outcome at varying prevalences of poor outcomes

100 100
98.4 97.5 96.6
90 PPV 90
NPV
80 80

70 70
Predictive value (%)

60 60

50 50

40 42.9 40

30 34.6 30

20 25.0 20

10 10

0 0
0 5 15 20 25 30

Prevalence of poor outcome (%)

3.3.2 Tests to identify the risk of a poor treatment outcome during TB treatment
Assuming that the prevalence of poor outcomes during treatment is 10%4 and assuming minimal
TPP accuracy targets – that is, 75% sensitivity and 80% specificity for a poor treatment outcome –
the test would have an NPV of 96.6% and a PPV of 29.4%. Therefore, in the hypothetical cohort of
1 000 people with TB, 25 of the 100 with a poor response to treatment would be missed, and 180 of
the 900 people with a good response to treatment would be incorrectly identified as having a poor
response and, therefore, might be overtreated. Under optimal TPP accuracy targets of ≥90% sensitivity
and specificity, only 9 of 100 people would be missed, and 90 of 900 would be overtreated (Fig. 4).

Fig. 4. Negative predictive values (NPVs) and positive predictive values (PPVs) for a
test with 90% sensitivity and 90% specificity for detecting those at risk of a poor
TB treatment outcome at varying prevalences of poor outcomes

100 100
98.8 98.1 97.3
90 PPV 90
NPV
80 80

70 70
Predictive value (%)

69.2
60 60
61.4
50 50
50
40 40

30 30

20 20

10 10

0 0
0 5 10 15 20 25 30

Prevalence of poor outcome (%)

4
The prevalence of poor outcomes is assumed to decrease during and at the end of treatment because treatment would already have
failed some patients or they would have had another unfavourable outcome.

22 Target product profiles for tests for tuberculosis treatment monitoring and optimization
3.3.3 Tests to detect people with poor outcomes at the end of TB treatment

Assuming a 5%4 prevalence of poor treatment outcomes by the end of treatment, including
early relapse, a test with 80% sensitivity and 90% specificity (i.e. the TPP minimal accuracy targets)
for detecting poor outcomes would have an NPV of 98.8%, and a PPV of 29.6% (Fig. 5). Therefore,
in the hypothetical cohort of 1 000 TB people, 10 of the 50 with a poor treatment outcome would
be missed, and 95 of the 950 people with a good treatment response would be incorrectly identified as
having a poor response and, therefore, might be overtreated. Under optimal accuracy targets of ≥95%
sensitivity and ≥95% specificity, only 5 out of 100 people would be missed and 48 of 950 overtreated.

Fig. 5. Negative predictive values (NPVs) and positive predictive values (PPVs) for a
test with 80% sensitivity and 90% specificity for detecting a poor outcome at
varying prevalences of poor TB treatment outcomes

100 100
98.8 97.6 96.2
90 90

80 80

70 70
Predictive value (%)

60 60
58.5
50 50
47.1
40 40

30 30
29.6
PPV
20 NPV 20

10 10

0 0
0 5 10 15 20 25 30

Prevalence of poor outcome (%)

3.4 Costs and cost effectiveness

The cost–effectiveness modelling included in these TPPs has shown that tests used during treatment
that meet the minimal targets (i.e. ≥75% sensitivity, ≥ 80% specificity) or optimal targets (≥90%
sensitivity, ≥90% specificity) for detecting people with an inadequate response to treatment could
considerably reduce unfavourable treatment outcomes and associated DALYs. A test fulfilling optimal
targets would yield the highest health impact by optimizing treatment and preventing poor outcomes
in more people as a result of better accuracy or by reaching more people (e.g. people who have scarce
sputum or have extrapulmonary TB), or a combination of these. Based on modelling, it was estimated
that 22% (uncertainty interval: 16% to 29%) of unfavourable treatment outcomes occurring in
the absence of treatment monitoring could be averted through tests that meets the TPP optimal
target criteria.

Currently, the tests most commonly used by TB programmes for monitoring TB treatment are sputum
smear microscopy and, sometimes, sputum culture or chest X-ray. These range in cost from US$
3.00 to US$ 20.00. Based on the cost–effectiveness modelling, tests meeting the optimal criteria for
sensitivity and specificity to detect a poor response to treatment should cost less than US$ 25.00
to achieve better health impacts at lower costs compared with sputum smear microscopy (Fig. 6).
The acceptable costs of tests may depend on how well they meet or surpass the targets described
in these TPPs.

3. Target product profiles 23

However, tests must also be affordable for TB programmes, as tests that are not affordable are often
not implemented or have poor market penetration, even if they are cost effective. Costs should also be
compatible with ensuring wide and equitable access and scale up, and affordability is essential for uptake
in lower-income countries. Furthermore, developers should be aware that costs, resource requirements,
and issues of equity, acceptability and feasibility are important criteria used by WHO to recommend
technologies such as the tests specified in these TPPs (31). Manufacturers should ensure that tests are
quality assured, widely available in a timely fashion and supplied in sufficient quantities to meet needs
of the affected populations. The required WHO quality standard is prequalification or similar certification
from a stringent regulatory agency or WHO Listed Authority (32).

The cost of less than US$ 25.00 is indicative only and allows for different pricing. Novel tests for
monitoring and optimizing TB treatment should be priced so that their use and implementation are
cost effective given the reduction in costs and morbidity associated with identifying people at risk
of poor outcomes from TB treatment. It is expected that public support for the development of the
tests and market shaping would lead to lower test costs.

Fig. 6. Average cost per treatment monitoring test performed in relation to

willingness to pay per additional disability-adjusted life year (DALY) averted relative
to testing with sputum smear microscopya
80
US$ 1575.00 per DALY averted
using smear microscopy
70
Average cost of monitoring test (US$)

60

50

40

30
US$ 24.75

20

10

0
1000 2000 3000 4000 5000
Willingness to pay per additional DALY averted (US$)

a
The figure shows estimates for a hypothetical, non-sputum-based test for TB treatment monitoring with optimal sensitivity
and specificity, according to criteria in the target product profiles. At a willingness-to-pay threshold of US$ 1 575 per additional
DALY averted (the best estimate of cost per DALY averted under smear microscopy–based testing), the test should cost ≤
US$ 24.75. The purple line indicates the best estimate; the grey shaded area indicates the 95% uncertainty interval.

There are additional cost considerations for the different use cases addressed in these TPPs. A test
that is done at the start of treatment and can direct people to optimal treatment early is likely to lead
to cost savings (e.g. shorter treatment regimens) and these can be reflected in the cost of the test.
Using serial testing – that is, monitoring tests conducted during consecutive months of treatment – to
identify individuals with an inadequate response to treatment may be advisable if the test is affordable.
Modelling suggests that serial testing should start during the early phase of treatment (Fig. 7). Serial

24 Target product profiles for tests for tuberculosis treatment monitoring and optimization
testing strategies may also provide opportunities to detect trends in treatment response over time and to
follow up on people for whom treatment was modified to ensure that the modified regimen is beneficial.

Ideally, tests will not require instruments, but if instruments are required these should be affordable
for TB and other health programmes, as capital costs are often a barrier to implementation. Higher
instrument costs may be more acceptable for multiuse platforms, but they would have to be supported
by evidence of their cost effectiveness. Instrument costs should include warranties, service contracts
and technical support for ≥3 years. Different models of instrument provision should be considered,
such as rental contracts or cost-per-result models.

Fig. 7. Model projections of costs per unfavourable treatment outcome averted for different
strategies of single time point and serial testing (i.e. repeated monitoring for
consecutive months) during TB treatment using a hypothetical sputum-based test that meets
optimal target criteriaa
3.3
Monitoring started
during month
1
3.2 2
3
Average cost / unfavorable treatment
outcome averted (thousand US$)

3.1

3.0

2.9

2.8

2.7

2.6

2.5
1 2 3 4 5 6

Month of treatment

a
The figure shows that the cost per unfavourable treatment outcome averted can be lower for serial testing compared with
single time point testing if it is started early in treatment, and that the cost is higher in later stages of treatment

3.5 Prioritization of test characteristics

It might be difficult to develop tests that adhere to all of the targets listed in the TPPs. In this light,
the Scientific TPP Development Group considered which characteristics should be prioritized when
developing and assessing potential tests for monitoring and optimizing TB treatment. The process
of prioritizing characteristics was also informed by the stakeholder survey, which asked specific
questions about priorities for characteristics, patient care pathway mapping and characteristics that
were considered barriers to implementing current monitoring tools. The following were considered
priority characteristics:
• diagnostic accuracy (sensitivity and specificity)
• time to result
• sample type
• target placement of test
• cost.

3. Target product profiles 25

References

1. Meeting report of the WHO expert consultation on drug-resistant tuberculosis treatment

outcome definitions, 17–19 November 2020. Geneva: World Health Organization; 2021 (https://
apps.who.int/iris/handle/10665/340284, accessed 8 June 2023).
2. Stadler JAM. Updated WHO definitions for tuberculosis outcomes: simplified, unified and
future-proofed. Afr J Thorac Crit Care Med. 2022;28:10.7196/AJTCCM.2022.v28i2.224.
doi:10.7196/AJTCCM.2022.v28i2.224.
3. Global tuberculosis report 2022. Geneva: World Health Organization; 2022 (https://
apps.who.int/iris/handle/10665/363752, accessed 8 June 2023).
4. WHO consolidated guidelines on tuberculosis. Module 4: treatment – tuberculosis care and support.
Geneva: World Health Organization; 2020 (https://apps.who.int/iris/handle/10665/353399, accessed
8 June 2023).
5. WHO operational handbook on tuberculosis. Module 4: treatment – drug-resistant
tuberculosis treatment, 2022 update. Geneva: World Health Organization; 2022 (https://
apps.who.int/iris/handle/10665/365333, accessed 8 June 2023).
6. WHO consolidated guidelines on tuberculosis. Module 4: treatment – drug-susceptible
tuberculosis treatment. Geneva: World Health Organization; 2022 (https://
www.who.int/publications/i/item/9789240048126, accessed 8 June 2023).
7. Horne DJ, Royce SE, Gooze L, Narita M, Hopewell PC, Nahid P, et al. Sputum monitoring during
tuberculosis treatment for predicting outcome: systematic review and meta-analysis. Lancet Infect
Dis. 2010;10: 387–94. doi:10.1016/S1473–3099(10)70071–2.
8. Friedrich SO, Rachow A, Saathoff E, Singh K, Mangu CD, Dawson R, et al. Assessment of the
sensitivity and specificity of Xpert MTB/RIF assay as an early sputum biomarker of response to
tuberculosis treatment. Lancet Respir Med. 2013;1:462–70. doi:10.1016​/S2213–2600(13)70119-X.
9. Nakaggwa P, Odeke R, Kirenga BJ, Bloss E. Incomplete sputum smear microscopy monitoring
among smear-positive tuberculosis patients in Uganda. Int J Tuberc Lung Dis. 2016;20:594–99.
doi:10.5588/IJTLD.15.0591.
10. Izudi J, Tamwesigire IK, Bajunirwe F. Treatment supporters and level of health facility influence
completion of sputum smear monitoring among tuberculosis patients in rural Uganda: a mixed-
methods study. Int J Infect Dis. 2020;91:149–55. doi:10.1016/j.ijid​.2019.12.003.
11. Jo KW, Yoo JW, Hong Y, Lee JS, Lee SD, Kim WS, et al. Risk factors for 1-year relapse of
pulmonary tuberculosis treated with a 6-month daily regimen. Respir Med. 2014;108:654–59.
doi:10.1016/J.RMED.2014.01.010.
12. Mitchison DA. Assessment of new sterilizing drugs for treating pulmonary tuberculosis by culture
at 2 months. Am Rev Respir Dis. 1993;147:1062–3. doi:10.1164/AJRCCM/147.4.1062.
13. Nahid P, Dorman SE, Alipanah N, Barry PM, Brozek JL, Cattamanchi A, et al. Official American Thoracic
Society/Centers for Disease Control and Prevention/Infectious Diseases Society of America Clinical
Practice Guidelines: treatment of drug-susceptible tuberculosis. Clin Infect Dis. 2016;163:e147–95.
doi:10.1093/cid/ciw376.

References 27
14. Migliori GB, Tiberi S, Zumla A, Petersen E, Chakaya JM, Wejse C, et al. MDR/XDR-TB management
of patients and contacts: challenges facing the new decade. The 2020 clinical update by the Global
Tuberculosis Network. Int J Infect Dis. 2020;92S:S15–25. doi:10.1016/J.IJID.2020.01.042.
15. Imperial MZ, Nahid P, Phillips PPJ, Davies GR, Fielding K, Hanna D, et al. A patient-level pooled
analysis of treatment-shortening regimens for drug-susceptible pulmonary tuberculosis. Nat Med.
2018;24:1708–15. doi:10.1038/S41591–018–0224–2.
16. Turkova A, Wills GH, Wobudeya E, Chabala C, Palmer M, Kinikar A, et al. Shorter treatment
for nonsevere tuberculosis in African and Indian children. New Engl J Med. 2022;386:911–22.
doi:10.1056/NEJMoa2104535.
17. WHO consolidated guidelines on tuberculosis. Module 5: management of tuberculosis
in children and adolescents. Geneva: World Health Organization; 2022 (https://
apps.who.int/iris/handle/10665/352522, accessed 5 February 2023).
18. Paton NI, Cousins C, Suresh C, Burhan E, Chew KL, Dalay VB, et al. Treatment strategy for rifampin-
susceptible tuberculosis. N Engl J Med. 2023;388:873–87. doi:10.1056/NEJMoa2212537.
19. Target regimen profiles for tuberculosis treatment. 2023 edition. Geneva: World Health Organization;
2023.
20. Papineni P, Phillips P, Lu Q, Cheung YB, Nunn A, Paton N. TRUNCATE-TB: an innovative trial design
for drug-sensitive tuberculosis. Int J Infect Dis. 2016;45:404. doi:10.1016/J.IJID.2016.02.863.
21. Heyckendorf J, Georghiou SB, Frahm N, Heinrich N, Kontsevaya I, Reimann M, et al. Tuberculosis
treatment monitoring and outcome measures: new interest and new strategies. Clin Microbiol Rev.
2022;35:e0022721. doi:10.1128/CMR.00227–21.
22. Zimmer AJ, Lainati F, Aguilera Vasquez N, Chedid C, McGrath S, Benedetti A, et al. Biomarkers that
correlate with active pulmonary tuberculosis treatment response: a systematic review and meta-
analysis. J Clin Microbiol. 2022;60:e0185921. doi:10.1128/JCM.01859–21.
23. Mapping care pathways. London: National Institute for Health and Care Excellence;
2015 (https://www.nice.org.uk/media/default/About/what-we-do/Into-practice/HTAP/
HTAPMappingCarePathwaysResource.pdf, accessed 8 June 2023).
24. Micocci M, Gordon AL, Allen AJ, Hicks T, Kierkegaard P, Mclister A, et al. COVID-19 testing in English
care homes and implications for staff and residents. Age Ageing. 2021;50:668–72. doi:10.1093/
ageing/afab015.
25. Hills NK, Lyimo J, Nahid P, Savic RM, Lienhardt C, Phillips PPJ. A systematic review of endpoint
definitions in late phase pulmonary tuberculosis therapeutic trials. Trials. 2021;22:515. doi:10.1186/
S13063–021–05388–1.
26. Horne DJ, Royce S, Gooze L, Narita M, Hopewell PC, Nahid P, Steingart KR. Sputum monitoring
during tuberculosis treatment for predicting outcome: systematic review and meta-analysis. Lancet
Infect Dis 2010;10(6):387–94. doi:10.1016/S1473–3099(10)70071–2.
27. Davis JL, Cattamanchi A, Cuevas LE, Hopewell PC, Steingart KR. Diagnostic accuracy of same-day
microscopy versus standard microscopy for pulmonary tuberculosis: a systematic review and meta-
analysis. Lancet Infect Dis 2013; 13(2):147–54. doi:10.1016/S1473–3099(12)70232–3.
28. Target product profile for next-generation TB drug-susceptibility testing at peripheral centres. Geneva:
World Health Organization; 2021 (https://apps.who.int/iris/handle/10665/343656, accessed 8 June
2023).
29. Peetluk LS, Ridolfi FM, Rebeiro PF, Liu D, Rolla VC, Sterling TR. Systematic review of prediction
models for pulmonary tuberculosis treatment outcomes in adults. BMJ Open. 2021;11:e044687.
doi:10.1136/BMJOPEN-2020–044687.

28 Target product profiles for tests for tuberculosis treatment monitoring and optimization
30. Imperial MZ, Phillips PPJ, Nahid P, Savic RM. Precision-enhancing risk stratification tools for selecting
optimal treatment durations in tuberculosis clinical trials. American J Respir Crit Care Med.
2021;204:1086–96. doi:10.1164/RCCM.202101–0117OC.
31. Moberg J, Oxman AD, Rosenbaum S, Schünemann HJ, Guyatt G, Flottorp S, et al. The GRADE
Evidence to Decision (EtD) framework for health system and public health decisions. Health Res
Policy Syst. 2018;16:45. doi:10.1186/s12961–018–0320–2.
32. Evaluating and publicly designating regulatory authorities as WHO listed authorities: policy document.
Geneva: World Health Organization; 2021 (https://apps.who.int/iris/handle/10665/341749, accessed
8 June 2023).

References 29
Annexes

Annex 1. Declarations of interests

The following members of the Scientific TPP Development Group declared no interests that could
conflict with the objectives of the Target Product Profiles: Abdulkadir Civan, Frank Cobelens,
Mustapha Gidado, Ankur Gupta-Wright, Rumina Hasan, Cathy Hewison, Kobto Koura, Mikashmi Kohli,
Christian Lienhardt, Patrick Lungu, Emily MacLean, Florian Marx and Lindsay McKenna.

The following members of the development group declared interests that were judged not to conflict
with the objectives of the Target Product Profiles:

Daniela Cirillo declared research support to the Ospedale San Raffaele of US$ 38 600 from the
TB Alliance in an unrestricted grant for a multipartner test of minimum inhibitory concentrations
for pretomanid, which ended in 2020, and a grant of US$ 62 629 from the European Committee on
Antimicrobial Susceptibility Testing (EUCAST) to coordinate work on a standard protocol that involved
reference laboratories for different anti-TB medicines.

Claudia Denkinger declared that during 2014–2019 she was head of the Tuberculosis Programme
at the FIND, a non-profit global health organization based in Geneva, Switzerland. At FIND she was
involved in exploratory work on TPP development and tools for treatment monitoring. FIND never
received industry funding for this purpose. In her current role, she continues to work on TB diagnostics,
specifically on research related to treatment monitoring tools, but she has never received industry
funding for this purpose.

Stephen Gillespie declared research grants to the University of St Andrews from the European Union
and the European and Developing Countries Clinical Trials Partnership (EDCTP) of € 1 million over
5 years and noted that his group has received payments to support testing of the LifeArc TB-MBLA
(Mycobacterium tuberculosis molecular bacterial load assay) kit (< £ 10 000). His group has received
grants that have been used to test the utility of the TB-MBLA kit as noted above. The University has
registered the trademark VitalBacteria to allow them to sell test kits for research use at cost. The
group has knowledge related to TB-MBLA, and a patent filing with regard to some aspects of the
work is in process.

Delia Goletti declared receiving honoraria for public lectures from bioMérieux (€ 1 500 in 2021) and
QIAGEN (€ 2 500 for 2021 and 2022), as well as honoraria for contributing to the development of
tests (US$ 122 500 from Quidel for 2019, 2020 and 2021; € 650 from PBD Biotech in 2022). She
also declared that the National Institute for Infectious Diseases L. Spallanzani in Rome, Italy, where
she works, received grants to evaluate new tests for diagnosing TB infections from bioMérieux
(€45 000 for each of two grants).

Annexes 31
Anneke Hesseling declared research support to Stellenbosch University in the form of US$ 4 million
per annum for several investigator-initiated research studies, with the US National Institutes of
Health (NIH)–funded IMPAACT network, the Tuberculosis Trials Consortium, Unitaid, Biomedical
Research Computing–Wellcome Trust, the South African Medical Research Council and the EDCTP.

Timothy McHugh declared a consultancy with Lumora Ltd (Erba Molecular) (£ 2 000, ceased in 2019),
a research collaboration with the TB Alliance (worth approximately £ 2 000 000 during 2016–2022), as
well as a research award received from the European Union’s Innovative Medicines Initiative UNITE4TB
project (€ 1 000 000, current).

Morten Ruhwald declared previous employment at the Statens Serum Institut and notes that employees
and former employees can be paid up to € 40 000 in taxable income if a license agreement involving
patents with the employee as an inventor generates an extraordinarily high income for the Statens
Serum Institut.

Rada Savic declared research funding or grants to her employer from the Bill & Melinda Gates
Foundation (BMGF), Critical Path Institute/BMGF, CZ BioHub Unitaid, the Global Alliance for TB
Drug Development, Innovative Medicines Initiative, US NIH, National Institute of Allergy and Infectious
Diseases (NIAID, part of the NIH), NIH/NIAID/Johns Hopkins University (JHU), NIH / NIAID / Rutgers, NIH/
NIAID/Stellenbosch, and WHO for a value of US$ 18.5 million (2016–2028), as well as a leadership
grant from the AIDS Clinical Trials Group.

Thomas Scriba declared research support including grants from the US NIH, the Bill & Melinda
Gates Foundation, the South African Medical Research Council and the EDCTP to his employer in
addition to a patent held by the University of Cape Town.

32 Target product profiles for tests for tuberculosis treatment monitoring and optimization
Annex 2. Results of the stakeholder consultation and Delphi survey

1. Stakeholder consultation
At the start of the target product profile (TPP) development process, stakeholders were surveyed to
gain input about their needs as they related to the TPPs.

A total of 45 individuals responded, including researchers, scientists and clinicians; staff at national
TB programmes and ministries of health, nongovernmental organizations, laboratories, public health
agencies and implementing partners; as well as developers of diagnostics, biomarkers and tests for TB;
and representatives of civil society organizations. Respondents came from all WHO regions except
for the Eastern Mediterranean. Altogether 35 (78%) respondents had worked in TB for more than
10 years.

The first set of questions in the survey focused on the proposal to examine three distinct use
case scenarios: a test to monitor treatment (i.e. during treatment), a test for cure (i.e. at the end
of treatment) and a test for disease severity (i.e. at the start of treatment). Respondents were asked
to indicate their agreement with definitions proposed for each of the tests. There was 84–100%
agreement (38–45 respondents) with the proposed TPP categories and definitions, but comments
highlighted a preference for programmes to have one test that could be used in all three scenarios.
In the same spirit, it was proposed that one TPP document should be developed to cover the three
scenarios instead of three separate TPPs.

When asked to prioritize the goals that should guide the development of the TPPs, the highest priority
was given to identifying people who would benefit from a change in their TB treatment, followed by
identifying people who have had a poor bacteriological response to treatment and identifying those
with an adequate response to treatment. The most important characteristic was considered to be
diagnostic accuracy, followed by the turnaround time for results and the setting in which the test
could be used. In addition, participants suggested that the acceptability of the test by people with
TB was important and indicated that a simple self-test could be advantageous.

Most participants considered it important that the TPPs include recommendations for special
populations (76% agreement, 35 respondents) so that their needs are specifically considered. The
most important population groups in order of priority were children, people living with HIV, people
with drug-resistant TB, pregnant and lactating women, and people with extrapulmonary TB.

There was general agreement (82%, 37 respondents) that it is important to be able to monitor treatment
using samples other than sputum, and the preferred sample types in order of priority were saliva,
urine, blood and imaging, followed by monitoring based on clinical features or measurements alone.

The ideal times to conduct a test to monitor treatment in order of priority were at 1 month after
starting treatment, within days to 1–2 weeks of starting treatment and if a patient was experiencing
clinical deterioration. The preferred frequencies for tests used to monitor treatment were in order
of priority, at the start of treatment and then again when a decision about treatment needs to
be made, and multiple times after starting treatment, followed by only once, when a decision about
treatment needs to be made. The ideal timing for conducting a test of cure was in order of priority,
at the end of TB treatment, 1 month prior to completing treatment and after completion to check
for early relapse, followed by early during TB treatment.

Annexes 33
2. Delphi survey
Participants of the technical consultation were sent the draft v. 0.0 TPP and a Delphi survey, as an
integral part of the consultation process. The survey was sent to 47 people of whom 29 (62%)
submitted a complete response. Participants were asked to express their level of agreement with
the proposed targets according to a predefined Likert scale ranging from 1 to 5 (1 – agree, 2 –
somewhat agree, 3 – neither agree nor disagree, 4 – somewhat disagree and 5 – disagree). Individuals
were asked to provide comments or alternative targets when they did not agree with a proposed
target (i.e. those scored at 3, 4 or 5). The targets on which fewer than 80% of participants agreed
were discussed further at the virtual technical consultation.

There was general agreement on the targets for assay or instrument design, target placement of test,
target user of test, maintenance, quality control, sample throughput, target population, sensitivity
and specificity. Fewer than 80% of participants (68%) agreed on the proposed minimal target for time
to result of 3 days or less and this was changed to 1 day or less in the subsequent discussion. There was
also disagreement (only 65% agreed) on the minimal requirement for sample type which was initially
proposed to be sputum only. After further discussion this was changed to “for use with a minimally
invasive respiratory sample (i.e. not limited to sputum alone)”. The Delphi survey also showed that
fewer than 80% of participants agreed (79%) on the optimal target proposed for timing for tests used
at treatment initiation to identify people with TB who require a more intensive treatment regimen.
In the subsequent discussions this target was adapted from “≤ 3 days of starting treatment” to “up
to 7 days after starting treatment”. Although the Delphi survey indicated a low level of agreement
(76%) on the proposed target for frequency for tests used to identify people at risk of a poor
treatment outcome during TB treatment, after further discussion in the meeting participants agreed
to maintain this target as “for use once at follow-up visit during treatment (without the need for
baseline measurement)”.

34 Target product profiles for tests for tuberculosis treatment monitoring and optimization
Annex 3. Technical consultation for the development of target product
profiles for tests and biomarkers for monitoring and optimizing
tuberculosis treatment, 26–28 September 2022 (virtual meeting)

Agenda
Day 1: Monday, 26 September Chair: Daniela Cirillo
14.00–14.15 Welcome and opening remarks Tereza Kasaeva
14.15–14.30 Introduction and background to the meeting, TPP Saskia den Boon
development process and meeting agenda
14.30–14.45 Limitations of current tests and overview of tests in the Emily MacLean
pipeline: results from a systematic review
14.45–15.30 Draft TPPs and results of Delphi survey Ankur Gupta-Wright
15.30–15.45 Break
15:45–17:00 Discussion of TPP categories and attributes: All
sensitivity and specificity Introduction by Ankur
Gupta-Wright

Day 2: Tuesday, 27 September Chair: Morten Ruhwald

14.00–14.20 Patient care pathways Ankur Gupta-Wright
14.20–15.00 Country data and experiences
China Yuhong Liu and Hui Xia
India Sanjay Mattoo
Brazil Kleydson Andrade
15:00–15.15 Q&A and discussion about patient care pathways All
15.15–15.30 Group photo and break
15.30–16.00 Cost–effectiveness modelling: plans and inputs Florian Marx
16.00–16.15 Q&A and discussion about cost–effectiveness modelling All
16.15–17:00 Discussion about TPP categories and attributes: sample type All
and target population
Introduction by Ankur
Gupta-Wright

Day 3: Wednesday, 28 September Chair: Frank Cobelens

14.00–14.05 Use case scenarios Claudia Denkinger
14.05– 14.15 Briefing on the development of WHO target regimen Fuad Mirzayev
profiles for TB treatment
14.15–14.45 Study design guidance Emily MacLean
14.45–15.00 Q&A and discussion about study design guidance All
15.00–15.15 Break
15.15–16.30 Discussion about TPP categories and attributes: All
turnaround time, timing and frequency
Introduction by Ankur
Gupta-Wright
16.30–16.50 Recap and summary Claudia Denkinger
16.50–17.00 Next steps and closing Saskia den Boon and Matteo
Zignol

Annexes 35
Participants
Macarthur Charles, US Centers for Disease Control and Prevention, Atlanta, Georgia, USA
Daniela Cirillo*, San Raffaele Institute, Milan, Italy
Abdulkadir Civan, University of Heidelberg, İzmir, Türkiye
Frank Cobelens*, Amsterdam Institute for Global Health and Development, Amsterdam, Netherlands
(Kingdom of the)
Claudia Denkinger*, University of Heidelberg, Heidelberg, Germany
Keertan Dheda, University of Cape Town and London School of Hygiene and Tropical Medicine, Cape
Town, South Africa
Norbert Djeka, National TB Programme, Pretoria, South Africa
Kathy Eisenach, independent consultant, Little Rock, Arkansas, USA
Ronald Allan Fabella, Disease Prevention and Control Bureau, Department of Health, Manila, Philippines
Mustapha Gidado, KNCV TB Plus, The Hague, Netherlands (Kingdom of the)
Stephen Gillespie*, University of St Andrews, St Andrews, United Kingdom
Delia Goletti, Translational Research Unit, National Institute for Infectious Diseases–Scientific Institute
for Research, Hospitalization and Healthcare (INMI-IRCCS), Rome, Italy
Ankur Gupta-Wright*, University College London, United Kingdom, and University of Heidelberg,
Heidelberg, Germany
Rumina Hasan, Aga Khan University, Supranational Reference Laboratory Karachi, Pakistan, and
London School of Hygiene and Tropical Medicine, United Kingdom
Anneke Hesseling, Stellenbosch University, Cape Town, South Africa
Cathy Hewison, Médecins Sans Frontières (MSF), Paris, France
Kobto Koura, The International Union Against Tuberculosis and Lung Disease, Paris, France
Mikashmi Kohli*, FIND, Geneva, Switzerland
Ravinder Kumar, Central TB Division, National Tuberculosis Elimination Programme, Delhi, India
Jose Lapa e Silva, Ministry of Health, Rio de Janeiro, Brazil
Christian Lienhardt, French National Research Institute for Sustainable Development, Montpellier,
France and FAST-TB Initiative, Civilian Research and Development Foundation Global, Arlington,
Virginia, USA
Yuhong Liu, Beijing Chest Hospital, Beijing, China
Emily MacLean*, University of Sydney, Sydney, Australia
Sanjay Kumar Mattoo, Central TB Division, National Tuberculosis Elimination Programme, Delhi, India
Florian Marx, University of Heidelberg, Berlin, Germany
Timothy McHugh, University College London, London, United Kingdom
Lindsay McKenna, Treatment Action Group, New York, New York, USA
Morten Ruhwald*, FIND, Geneva, Switzerland
Rada Savic*, University of California San Francisco, San Francisco, California, USA
Thomas Scriba, University of Cape Town, Cape Town, South Africa
Christine Sekaggya-Wilthsire, Infectious Diseases Institute, Kampala, Uganda
Boitumelo Semete-Makokotlela, South African Health Products Regulatory Authority, Pretoria, South
Africa
Kelly Stinson, Cultura, LCC, Atlanta, Georgia, USA

36 Target product profiles for tests for tuberculosis treatment monitoring and optimization
Ezio Tavora dos Santos Filho, WHO Civil Society Task Force and Rio de Janeiro Federal University, Rio
de Janeiro, Brazil
Nguyen Thuy Thuong, Oxford University Clinical Research Unit, Ho Chi Minh City, Viet Nam
Cesar Ugarte-Gil, Universidad Peruana Cayetano Heredia, Lima, Peru
Hui Xia, National Center for TB Control and Prevention, China Center for Disease Control, Beijing,
China
* member of the task force

Funding agencies
Sevim Ahmedov, United States Agency for International Development, Washington, DC, USA
Grania Brigden, The Global Fund to Fight AIDS, Tuberculosis and Malaria, Geneva, Switzerland
Debra Hanna, Bill & Melinda Gates Foundation, Seattle, Washington, USA

Commercial developers of tests for TB treatment monitoring and optimization

Devasena Gnanashanmugam, Cepheid, Washington, DC, USA
Ammar Jagirdar, Qure.ai (former employee), Mumbai, India
Nakaishi Kazunari, Tauns Laboratories, Shizuoka, Japan
Ahmed Maged, Abbott, Chicago, Illinois, USA
Megumi Komada, LSI Medience, Tokyo, Japan
Jerome Nigou, Institut de Pharmacologie et de Biologie Structurale, Toulouse, France
Akos Somoskovi, Roche, Pleasanton, California, USA
Sruti Sridhar, Qure.ai, Mumbai, India

World Health Organization, Global Tuberculosis Programme, Geneva, Switzerland

Saskia den Boon
Dennis Falzon
Nazir Ismail
Fuad Mirzayev
Samuel Schumacher
Kerry Viney
Matteo Zignol

World Health Organization

Kleydson Andrade, WHO country office in Brazil, Brasilia, Brazil
Corinne Merle, Special Programme for Research and Training in Tropical Diseases, Geneva, Switzerland
Nkateko Mkhondo, WHO country office in South Africa, Pretoria, South Africa
Ernesto Montoro, Pan American Health Organization, Washington, DC, USA
Kirankumar Rade, WHO country office in India, Delhi, India
Martin van den Boom, WHO Regional Office for the Eastern Mediterranean, Cairo, Egypt
Askar Yedilbayev, WHO Regional Office for Europe, Copenhagen, Denmark
Chen Zhongdan, WHO country office in China, Beijing, China

Annexes 37
Annex 4. Scientific TPP Development Group meeting, 27–29 March 2023,
Istanbul, Türkiye (hybrid meeting with remote connection)

Agenda
Day 1: Monday, 27 March Chairs: Ankur Gupta-Wright
and Mikashmi Kohl
9.30–10.00 Arrival and registration
10.00–10.10 Welcome and opening remarks Matteo Zignol
10.10–10.30 Meeting objectives, presentation of participants and review Dennis Falzon
of declarations of interest
10.30–10.50 TPP development process and meeting agenda Saskia den Boon
10:50–11.20 Break
11.20–11:50 Feedback from the public consultation Emily MacLean
11:50–12:30 Discussion All
12.30–13.30 Lunch
13.30– 14.00 Analysis of patient care pathways Ankur Gupta-Wright
14:00–15:00 Discussion All
15:00–15:30 Break
15:00– 17:30 Discussion and consensus-seeking about targets for test All
characteristics common to all TPPs

Day 2: Tuesday, 28 March Chairs: Daniela Cirillo and

Claudia Denkinger
9.00–9.45 Cost–effectiveness analysis Abdulkadir Civan and Florian
Marx
9:45–10:30 Discussion All
10.30–11.00 Break
11.00–12.00 Discussion and consensus-seeking on cost targets All
12.00–13.30 Group photo and lunch
13.30–15.30 Discussion and consensus-seeking about targets specific All
for tests to identify people for less intensive TB treatment
regimens at treatment initiation
15.30–16.00 Break
16.00–18:00 Discussion and consensus-seeking about targets specific for All
tests to identify poor responses to TB treatment

Day 3: Wednesday, 29 March Chairs: Emily MacLean and

Rada Savic
9.00–10.30 Discussion and consensus-seeking about targets specific for All
tests to identify people who can stop TB treatment

10.30–11.00 Break
11.00– 11.30 WHO target regimen profiles for TB treatment Fuad Mirzayev and Samuel
Schumacher
11.30–11.45 Recap and summary Claudia Denkinger
11.45–12.00 Next steps and closing Saskia den Boon and Matteo
Zignol
12.00–14.00 Lunch

38 Target product profiles for tests for tuberculosis treatment monitoring and optimization
Members of the Scientific TPP Development Group
Frank Cobelens*, Amsterdam Institute for Global Health and Development, Amsterdam, Netherlands
(Kingdom of the) (could not attend the meeting)
Daniela Cirillo*, San Raffaele Institute, Milan, Italy
Claudia Denkinger*, University of Heidelberg, Heidelberg, Germany
Mustapha Gidado, KNCV TB Plus, Netherlands (Kingdom of the)
Stephen Gillespie*, University of St Andrews, St Andrews, United Kingdom
Delia Goletti, Translational Research Unit, National Institute for Infectious Diseases–Scientific Institute
for Research, Hospitalization and Healthcare (INMI-IRCCS), Rome, Italy
Ankur Gupta-Wright*, University College London, United Kingdom, and University of Heidelberg,
Heidelberg, Germany
Rumina Hasan, Aga Khan University, Supranational Reference Laboratory, Karachi, Pakistan, and
London School of Hygiene and Tropical Medicine, United Kingdom
Cathy Hewison, Médecins Sans Frontières, Paris, France
Kobto Koura, The International Union Against Tuberculosis and Lung Disease, Paris, France
Mikashmi Kohli*, FIND, Geneva, Switzerland
Christian Lienhardt, French National Research Institute for Sustainable Development, Montpellier,
France, and and FAST-TB Initiative, CRDF (Civilian Research and Development Foundation) Global,
Arlington, Virginia, USA
Patrick Lungu, East, Central and Southern Africa Health Community, Lusaka, Zambia
Emily MacLean*, University of Sydney, Sydney, Australia
Timothy McHugh, University College London, London, United Kingdom
Lindsay McKenna, Treatment Action Group, New York, New York, USA
Morten Ruhwald*, FIND, Geneva, Switzerland (could not attend the meeting)
Rada Savic*, University of California San Francisco, San Francisco, California, USA
Thomas Scriba, University of Cape Town, Cape Town, South Africa
Christine Sekaggya-Wiltshire, Infectious Diseases Institute, Kampala, Uganda
* member of the task force

Funding agencies
Grania Brigden, The Global Fund to Fight AIDS, Tuberculosis and Malaria, Geneva, Switzerland
Debra Hanna, Bill & Melinda Gates Foundation, Seattle, Washington, USA
Cherise Scott, Unitaid, Geneva, Switzerland

Mathematical modellers
Abdulkadir Civan, University of Heidelberg, İzmir, Türkiye
Florian Marx, University of Heidelberg, Berlin, Germany

World Health Organization, Global Tuberculosis Programme, Geneva, Switzerland

Saskia den Boon
Dennis Falzon
Nazir Ismail
Fuad Mirzayev
Samuel Schumacher
Matteo Zignol
Annexes 39
For further information, please contact:
Global Tuberculosis Programme
World Health Organization
20, Avenue Appia CH-1211 Geneva 27 Switzerland
Web site: www.who.int/tb

42 Target product profiles for tests for tuberculosis treatment monitoring and optimization