PD Dr. med. Claudia Denkinger is the head of the Infectious Disease and Tropical Medicine Division at the University Hospital Heidelberg. In this edition of the GLOHRA Academy Series, she shared insights into her research on the development and evaluation of novel diagnostic tests for Tuberculosis, that are particularly relevant and suitable for low- and middle-income settings.
The GLOHRA Academy Series is a format of the Global Health Academy that presents and discusses global health research from the perspectives of the different research disciplines. Every first Tuesday in a month, a speaker shares insights of her/his research projects while focusing on research processes, methods as well as lessons learned. The Series directs special attention towards the added value of interdisciplinary and international cooperation in global health research. The Global Health Academy is an initiative of the German Alliance for Global Health Research (GLOHRA) with the explicit mission to support the next generation of global health researchers.
So transdisciplinary research to close the diagnostic Gap in tuberculosis care actually MEF already introduced Me Maybe One addition we’re at the um infectious disease and tropical medicine Institute here um in heidleberg um we have a growing infectious disease clinical footprint um and um uh growing like research Group which primarily focuses
Uh on diagnostics for Global Health uh but there are also other um people in the team that have more basic science translational aspects um or clinical uh uh research uh as well all right but today we’re talking about transdisciplinary research to close the diagnostic Gap in
Tuberculosis care and why do we care at all about that is the question and I hope the SL slide shows you the lack of access to diagnosis for tuberculosis is horrendous in fact we have 10 million about 10 million cases every year and of those about 41% never understand that they’re
Actually suffering uh from tuberculosis and um they are not notified uh in the case Registries if we look at that for children it’s even more horrendous in that the number numbers exceed like 65% actually never receiving a tuberculosis diagnosis this has gotten worse due to covid because people didn’t access care
Um which has resulted in as you can see here from this graph an increase in the number of deaths and this has been the case uh across many countries um and if you want to read further on this I always try to uh site the references at
The bottom so concrete rely why is this the case um so in tuberculosis care about 80% of people first present in community settings or Primary Care settings and there we only have a symptom screen u meaning you ask whether people have fever weight loss prolonged cough night sweats and you can already
Tell from those symptoms that these are very non-specific symptoms many uh diseases can give you a prolonged cough or um or or fever so at a place where most people present to care first we essentially don’t have a test we only have a poor symptom screen if people identified to be at
High risk they are referred to District of referral hospitals and then we have diagnostic tests such as microscopy which is actually not much better than what Robert Co had over 120 years ago and this is still the most widely used test globally despite the fact that the
Who recommends a PCR test since 2010 but that PCR test has the problem that it requires electricity temperature control and is costly and thus it doesn’t reach patients culture-based tests are uh few and far between a problem is that the culture in tuberculosis takes weeks so
Even if you have access to culture based tests which are the gold standard tests the result might come back so late that you never get the results because you already went back to your community and Primary Care Center or to your home um and the results never reaches you so
This is the breakdown why essentially we have this large diagnostic Gap tests don’t reach people and what we need is really a tool to screen and triage people recognize those that are at high risk of TB that’s rapid and easy to use and also we need to improve upon the
Test at the district and referral hospitals but there are um you first of all need to get the patients into the care pathway and the most important step for that is this if we think about what it takes to develop a novel diagnostic tests there are lots of obstacles confronted of by diagnostic
Innovators we call those The Valleys of death from concept to roll out of a test and first of all it’s in the in the early phases of scientific and Technical product design where a lot of the failures occur then there’s a lot of regulatory aspects and if you think
Where most of the test developers are they’re mostly in places when there’s no TV so actually access to um trial sites to where you can uh speak to the stakeholders in country and validate your novel test is a big hurdle and lastly planning for adoption supply chain evidence of impact is is another
Big hurdle and what we are trying to do here in the team is to essentially with our research across this diagnostic value cycle to address gaps and address these um valys of death that I mentioned so first of all we look at the need like what are is actually needed in the
Countries and how should it look like what would be acceptable usable um at those uh places so that we can um ensure that this technical design the scientific design of the product meets the needs of the people who are actually intending to use the test at the end
Then you often times have to make tradeoffs um because you have limited amount of funding or you want to get the highest impact with what you what you develop and so you could um assess the potential here mathematical modeling becomes relevant and then obviously you need to make sure you get the right
Targets develop the right test again here um you using the opinion of the of the the end users is of is a critical component in the comp conceptualization and realization of a tool and then you need to generate the data on accuracy clinical utility and value to inform wh
Policy or country-based policy and so with this we try to essentially um bring diagnostic tests for tuberculosis forward and you might ask why a scientific group academic group um or an NGO in some cases is trying to do that it’s simply because there’s hardly any money in diagnostic tests for Global
Health problems rather than focusing on cancer nobody would focus on a diagnostic test so we need to also support support diagnostic test developers bring products along to address these needs and make sure that they reach people so let’s go through this diagnostic value cycle and look at a
Little bit at the research um that is necessary there so I spoke about need and I also thought I introduce you a little bit to our team here so I put a picture in and the name of um the the team member who is focusing largely on
This so first of all looking at need for an accessible non-sputum B sample so I told you we don’t have a good test um in the communitybased settings or Primary Care settings and one big problem is that sputum is actually difficult to get imagine like I had a cold last week that
Was fine I was producing sputum no no big deal for me but imagine like a weak HIV a p weak person living with HIV in advanced stage weak enough not to be able able to produce butum or think of a child I told you the diagnostic Gap in
Children is so substantial think of a child being told can you bring up a sputum I tried that with my son who is sick six and he can now do it but like last year not yet so I think the cut off is about five six years when kids are
Able to cognitively process what it means to bring up a sputo um but so many people actually cannot produce a sputum and if you then think of um the availability of the sputum sample and the if you multiply it with a sensitivity of the test so the
Capability of a test to detect uh the disease um then actually for um expert which is the who recommended uh test uh for tuberculosis for expert in an HIV positive cohort you only get to diagnose about half of the people who have confirmed tuberculosis at the end on the
Other hand if you take a urine test that is actually only 50% sensitive so see here 90% 50% many more people can produce urine than people can produce sputum so actually you have almost the same number of people actually being diagnosed even though the sensitivity is much lower so
This is like an assessment in a study to that shows you you really need to go for non-sputum based samples if you want to reach all of these patients particularly in the community in Primary Care settings where you don’t have the capability to do an induced spum like
Stimulate the airway to to to produce a pum or do a gastric lavage where you stick a tube down the into the stomach which is often done for children to get a sample another um component to to understand need is to understand um um what is acceptable to people
And this is yeah I don’t expect you to read this text but it’s just like some work that we did with a medical student here Janis Herman um together with Partners in malavi and sambia where we spoke with patients healthcare workers test operators and decision making makers on a national level and
Essentially identified what would be barriers and facilitators uh to implement a test um and this was a very insightful we did that for a novel urine based test um and this is also applicable to urine based tests more more broadly um and has since then also
Been used to informs some who policy but we also more directly actually work uh with patients um and engage them in uh the communities that they first present for care and they that they would like to have diagnostic testing to happen in order to elicit their
Preferences and and here um one of my uh PhD students Maria delar Castro Nora um has worked a lot with um in in a large project which is happening right now in Zambia and Vietnam where we are trying to um get people to essentially trade describe their tradeoffs in a in a um
Discrete Choice experiment that’s an a method for qualitative uh research that actually has quantitative components and is um taken from economical um um not economical economics research um so you essentially ask people between options what they would prefer and a similar method a little bit uh less uh
Intense is best worth scaling which is um another research that we are doing also um in um Zambia Nigeria and the Philippines and all of this are to elicit essentially what would be most preferential for these uh uh these end users and patients um is it
Like for example like a urine based test a tongue swap is it um is it sputum or is it like an image so we’re looking at at what would be uh the best way to put our uh efforts in terms of development then as I had told you there
Are always trade-offs to be made and you should focus on what is really has the highest potential and this is actually quite um uh well feasible to do that by looking at mathematical modeling because um with mathematical modeling you can assess um the impact of a hypothetical
Test or a real test um for on a population basis both in respect to um identifying patients but also in respect to preventing deaths or transmission and so we are doing mathematical modeling in this case it’s a classical seir model um um that essentially follows the natural
History of TB um and has some um tweaks to reflect um uh special sub populations characteristics of the TB um and with this for example we look at uh currently uh what the additional impact would be of um uh of a test that reaches more people um
But might be less sensitive so we are able to weigh these trade-offs and this is work of Lucas brummer who is a medical student in my team um who actually looked at um what are cost Effectiveness thresholds so um the in the aspect of cost that I mentioned
Earlier I told you that the PCR test that the who recommends hasn’t got hasn’t gotten wide um acceptance because it’s quite costly so what would actually be acceptable what should it cost in order to to make a difference um and this is a different um uh screening considerations um with different tests
Essentially in different uh countries and uh what would be um most acceptable so now we have spoken about need about potential if I go back to this here now we are going onward to Discovery and here first of all you need to know what to go after right so we
Have now defined that actually um looking at novel sample types um like urine would probably be a good idea to uh because more people can produce it it’s more accessible but what would be um acceptable sample targets or targets for TB in those samples and this is work
That we did um together with a colleague at McGill now in Australia uh where we essentially mapped the the biomarker landscape of tuberculosis and looked at um what the performance was this is on a graph where you have sensitivity on the y axis specificity one minus specificity
On the x-axis so you test you want to actually have them up here in that corner and you can see that there are many studies U with biomarkers that fall into that um region that being said if you actually replicate the data do it with larger studies the bubble size is
The study size you see that many of those positive findings quickly drop off um and and uh those markers biomarkers are not so promising after all but um the other thing that you need to know is essentially what actually would be acceptable in terms of your characteristics of the test and this is
Work that the who does um to Define what actually would be the optimal Target product profile so how should the diagnostic test look like and I’m mentioning this here also because we have actually supported The Who on various occasions in developing these Target product profiles which was a lot
Of um stakeholder engagement but also modeling pathway care pathway mapping um and um and qualitative research in order to come to the characteristics such a Target product profile and also in order to Define what a test what a trial should look like to Define um um or to
To validate if a test meets those characteristics so we had said sputum not so good but other sample types could be um favorable um and so there I just wanted to quickly actually show some research on this so this was work work on urine uh which in part um I did
Already in my prior work life um at fine the foundation for Innovative new Diagnostics um where we worked with a company um to develop a better urine lamb test which is a urine-based TB Diagnostic and we’re able to improve substantially on it but it’s still not uh unfortunately on the market because
Um we discovered some issues in its stability and lot to lot variability but this was the initial trial and then the initial meta analysis that showed the promise of these tests where you could see that the sensitivity compared to a currently wh recommended test was doubled so there is potential in going
After a urine cell wall component of so cell wall component of tuberculosis in the urine um to develop better diagnostic tests and we are pursuing this further we also just got funding from Horizon Europe to look at bio aerosol so we all heard a lot about
Aerosol testing for for covid now we are trying to see whether this is actually feasible for tuberculosis diagnosis or possible for um establishing transmission risk um and this is work that we’ll be doing in the next five years um with a small Swiss startup trying to to uh develop easy sample um
Um methods um that could potentially improve the non-invasive sampling and get molecular tests closer to to the people and reach more people the work that we are doing in chora um addresses not the diagnoses but the screening component and I told you that this in my eyes is actually the most important
Component and here we are trying to essentially use data from large prevalence surveys uh globally to see whether we can do better than that four symptom screen that I told you earlier uh that was so um neither very sensitive nor very specific and so we have gathered essentially a prevalence
Surveys with the help of the wh from like um I have to count I think it was in 16 countries in total um um and also a lot of um trial data from various organizations um that we are now feeding into like a a global data set and you
Can imagine how much data cleaning and and um and work that was and we are now applying machine learning methods and also um basan uh latent class techniques in order to develop algorithms that we can then hopefully uh validate um and translate on to an easyto use um app um
Or digital Tool uh which first of all would help us Define what the risk of tuberculosis is in a in a given patient and then with an additional effort to Define um um what a treat test treatment threshold should be um to see what a treatment recommendation would result
From that so this is a very exciting project we’re doing this in clora for adults and we’re actually also wanting to to do this for for children where I told you the Gap is even larger so now we have a tool at hand imagine all of these Discovery efforts and
Development efforts were successful then we have to or we have a first prototype we have to validate them um and clinical validation is a big hurdle as I told you earlier because most of the companies that develop are here or here or here but the TB is here and so we have with
Um um NIH funding essentially developed this large TB trial Network um in 10 countries globally where we have uh evaluated over 40 prototypes to like more advanced versions of diagnostic tests already and are continuing to do so with um these Partners um and actually this all started during covid
Times we just had our first joint meeting imagine that’s been going on for three years this trial um we’re so exciting just like two weeks ago with all of the partners from around the world and so just to give you some examples we have done an accuracy trial
On oral swaps very easy to do we all did it uh in some form or the other for for covid and found that actually it would also be a very easily accessible sample type for TB not as good as sputum but if you reach more people you might accept
Trade-offs of sensitivity so this was a very promising study and as we get the results from these um from these uh um validation um trials we also obviously try to translate them um to uh to policy and this means that on the one hand we generate our evidence I told you on
Accuracy acceptability implementability we also do economic analysis this is the health Economist kader and my team we talk to other um um groups that are doing validation we actually established a joint steering committee with all of the major trial um consortia globally um to essentially um essentially bring all of
Our data together and I have to say I think our um trial network name and logo is the best ever um I’m very proud of that um but I’m happy that all of these people are finally coming together in order to advance TB Diagnostics and so how does polic translation work
Essentially it means weighing um evidence from different um areas with generated with different methods addressing different components like stakeholder values I told you qualitative work accuracy that’s the validation work Value Health economics or economic evaluation Equity of access this is something I didn’t speak about um because it’s more done with um um
With a model at this stage at later stage um you can also do it uh obviously with implementation studies and then um uh like what is actually the access how do people get access this is also something often done uh with modeling at this stage and so um I’m very much
Involved with the wh on these uh trans evidence translation to policy um as an expert advisor for TV Diagnostics that’s an elected um position for three years and we’ve been involved since 2019 on five uh recommendations and just to show you one example of a tradeoff this is a
Simple one we saw the circle is much more complex but there’s a test for TV screening actually that’s not so bad um it almost meets The Who Target product profile but it actually is quite costly and uh we figured that it was quite good in terms of its performance but it’s
Never going to reach scale because it’s a cost even if it were $2 um it would not be cost effective and the fact is that this cartridge is more kind of like five dollar and above so unlikely that we’re going to reach that scale or policy so this was just a quick run
Through to the um research um around the value cycle that we are doing for t tuberculosis I want to say that we have a lot of goals in tuberculosis I actually want to get um give diagnostic ownership to the TB to the patients presumed to have TB so that
They can um Define what their risk is and what they need to do and this really is done with self testing so this is my goal hopefully in my lifetime to make self testing for TB possible seems like obvious right we made it for covid possible why not for TB it’s much more
Complicated but we’re also looking Beyond TB other um disease areas and and questions and for covid like many others we did a lot um and just wanted to show that here and covid actually um has given as great opportunities just as much as it has given or a little bit U
Um but obviously disproportionate to the the um challenges it has given us but it hasn’t given us a lot of technical advances and this is exciting because we can now motivate these um diagnostic test developers to essentially translate their Innovation their technical Innovation to tuberculosis and this is
Work that we are doing in this large um uh Network together with Partners um trying to really Advance TB Diagnostics and get more developers into this field so with that um I just want to say two things one uh in my own interest I want to point out that we have um um
Infectious disease position open um for uh starting uh 1 of March for one year um and that we have a couple of research prisions open so if you’re excited about what I just told you think about it if you want to um find more information on
These um on these uh positions um email us or look on our website um and we will we’d be happy to tell you more this one requires German um for all these three uh it doesn’t require uh German for this one German would be beneficial because some of those clinical trials will
Actually happen in Germany and then the clor academy also asked me to um to say what I think Global health research is and I think from the what I shown you just now um I think I already told you but just to put it into words and I’m I took um um this
Definition from a paper in bmj global Health uh which I think is very opt uh it’s an area of research and practice committed to the application of multidisciplinary or transdisciplinary multi sectorial which transdisciplinary actually includes the sectoral aspect and culturally sensitive methods I spoke to you about needs
Knowing what actually is going to work for people what they want and approaches with a goal of reducing Health disparity that disparities that transcend National borders and with that I want to thank you for your attention and I hope I um yeah was able to tell you a little bit about tubercolosis
Diagnostics research or Global Health Diagnostics research what is relevant um and how we can work to address it and as said if you want to know more please visit our website