Good timing. Dr. Attia just posted an interview with Dr. Alex Aravanis about "...research and development of technologies and clinical tests utilizing the latest tools in DNA analysis and data science...Alex delves into two interconnected topics: liquid biopsies and epigenetics." I haven't listened yet, but if it's like all his other podcasts, it will be informative and detail oriented. (They both have degrees in engineering!) It might give you more "food" for thought on the topic to write about.
I coincidentally just listened to the podcast ready to share my thoughts in this comment section before seeing your comment. Dr. Attia and Dr. Aravanis provide the counter-arguments against Dr. Cifu's concerns about patient anxiety and overdiagnosis, however, I did not find their argument compelling. Please note I am just a medical student and still get tripped up on what makes a test effective (to my attending's frequent dismay..)
When questioned on whether using this test before we know it is effective at reducing cancer mortality or all-cause mortality (an end-point they basically say they find impossible to power to) is the right thing to do, Dr. Aravanis says the following (paraphrasing from 1hr 30mins if you want to check my work): If we have a test that can find early stage cancer, and a patient goes to an oncologist/onc surgeon with this information, those doctor with that information would treat that cancer. No one will say "oh the data isn't out yet on treating this based on this information, so I am going to let it be." So if the doctor is going to treat early cancer as long as they know it is there, why shouldn't we be finding all the early cancer we can find? Thereby, it is unethical not to just screen all the patients we can, whether or not we have run the study to determine if it is effective at reducing cancer mortality. (end paraphrase)
This seems to emphasize catching and treating cancers that are turtles and maybe some rabbits, but also finding birds just to find them as Dr. Prasad would put it. It is the optimistic conclusion that all data is good data. It is worrisome, but not surprising to think that the proponents of such a powerful tool do not understand that finding everything does not mean saving everyone, and that all patients will be happy to know about what is going on inside them, even if they can't do anything about it. As you said, Dr. Cifu, we may be harming patients physically, but more frequently mentally, in the process.
My optimism is that a test like this that looks for all cancers at once may actually be the first kind of screening tool that could clearly show all-cause mortality reduction rather than just cancer mortality reduction in a good trial. My pessimism is that they will not run the trial as they believe the mechanism over the evidence.
You make good points - your attending would be proud! :) I'll just push back gently on one thing you said, which was that not all patients would be happy to know what is going on inside them even if there isn't anything they can do about it. I think the point was that there's nothing much effective you can do when cancer is found late stage, but when you find it early, you actually CAN do something effective, and they can tell what type of cancer it is so they know how to treat it. But the main point I wanted to say is that Dr. Attia made clear early on that he is very up front with patients about the anxiety they might experience needlessly if they have a false positive, and also they should not consent unless they plan to do something further with a positive finding. Some of Attia's patients say, "No thanks, can't deal with that," and Attia is fine with that. In other words, he combines well informed consent with patient autonomy, and that would be my own preference. As someone who had a parent who died of cancer, I personally am more anxious WITHOUT screening. So I think it should be very much up to each patient based on their own personal and family history and preferences, leaving aside financial issues for now, as that's an entirely different basket of worms.
Really, though, someone needs to work faster on developing tests for silent killers like ovarian cancer. I've had three friends die of it. Can you blame a person for wanting SOME sensitivity, ANY sensitivity?
Specificity and sensitivity are not fixed properties of a test they have been shown to vary with prevalence and spectrum of the disease. In this case the population chosen to calculate the specificity, "healthy controls" should yield a specificity close to 100%. This wont be the case in practice. A similar test is being promoted in veterinary medicine for dogs. There is a lot of money in testing healthy individuals.
Findings In this systematic review and meta-analysis of 18 long-term randomized clinical trials involving 2.1 million individuals, colorectal cancer screening with sigmoidoscopy prolonged lifetime by 110 days, while fecal testing and mammography screening did not prolong life. An extension of 37 days was noted for prostate cancer screening with prostate-specific antigen testing and 107 days with lung cancer screening using computed tomography, but estimates are uncertain.
When it comes to screening, people completely lose the plot.
Screening ONLY works IF you can accurately detect a disease in the early stages AND treating that disease in the early stages alters the natural history and eventual outcome AND you don’t kill more people than you cure through false positives, unnecessary testing, etc.
The reason we have so many breast cancer survivors these days is that we are finding early tumours that probably never would have progressed to serious disease, but we are treating them all as if they would, while over-investigating like crazy! Big money maker for the mammogram machine makers, radiologists, pathologists, cancer society, etc. Clearly, from a business perspective, screening works if you can extract dollars from the patient’s wallet.
While I agree on the specific case of this test, I disagree with your background argument. I might suspect it depends on a general attitude towards how to practice medicine, but I'd rather avoid using claims no one made.
I don't understand the point that there's no evidence that early diagnosis improves outcomes.
First, the absence of evidence doesn't prove that there's no improvement. Second, it depends on how you act on these early diagnoses. You can do something stupid or even harmful to sell useless expensive stuff. Or you can perform life saving interventions, minimizing costs given the available resources. Early diagnosis doesn't seem to be the problem, unless you can prove that early diagnosis data is totally useless cost-benefit wise, for anything health related.
Then I don't understand the problem of overtreatment. Why an "honest" doctor would push overtreatment? The problem isn't diagnostics: it's "dishonesty" - however oversimplified is, using this concept to describe social phenomena that manifest in this context.
Findings In this systematic review and meta-analysis of 18 long-term randomized clinical trials involving 2.1 million individuals, colorectal cancer screening with sigmoidoscopy prolonged lifetime by 110 days, while fecal testing and mammography screening did not prolong life. An extension of 37 days was noted for prostate cancer screening with prostate-specific antigen testing and 107 days with lung cancer screening using computed tomography, but estimates are uncertain.
It doesn't surprise me and at the same time it doesn't sound like a real counter to my argument. Actually a study can only prove that a specific early diagnosis is useless conditioned to a specific medical follow up.
Let's abstract from a specific test and a specific diagnosis. I have an information: is it valuable? Is it useful towards any health related goal?
The answer is rarely "NO". Usually the answer is "it depends on the cost, the available intervention options, etc". A study (even a good meta-analysis) that shows that having an information didn't improve health outcomes (even the most important ones), solves only partly the medical conundrum. It just shows that either the information is useless OR we don't know what to do with it (yet).
That is to say: the study doesn't tell us if early diagnosis is indeed useless or not... Though it counters the marketing claims about the "usefulness" of specific examples of early diagnoses.
There is considerable empirical research, both published and a lot unpublished (in the hands of diagnostic companies) that show the absurdity, and later futility, of using "normal" patients in the identification and validation of diagnostic tests.
The simple way of describing this is that comparing a normal control to a patient with a condition is a comparatively easy comparison, rather than comparing patients for whom the differential includes the condition, but may be other conditions in the same organ, etc.
So a priori, a study that does prospectively take all comers does not represent real world medicine and data has shown will likely over estimate the diagnostic capabilities of the test. This will be the result for Galleri.
And as a corollary, if one wants to test everybody, not just those with conditions in the differential, the test will most likely just place patients into the differential rather than identify those with cancer.
The problem I have with any cancer detection is that if the test is positive, is it really trustworthy? Next, if it is then I am faced with radiation, chemo, surgery and endless drugs with not much hope of long term survival while wrecking my body further. No thanks. There are too many natural remedies to try that will at least be as effective without incurring more damage. Ask your oncologist about them and he will give you the "baffled' look. So reassuring.
With the current passion for surrogates, I will not be surprised if that becomes the standard of care for cancer screening one day with doctors being penalized for not ordering it. I’m also afraid of early cancer detection.
Thank you for the article. Could someone please explain to me how the pretest probability was calculated? Where did the 0.9% or 30 come from? Not doubting it, but I just didn't follow it. Thank you in advance.
Lifetime risk of pancreatic cancer about 1%. 90% of that is "regular" cancer in people over about 55. 30 is the # of years that someone is at risk of having new, treatable, cancer that is screen detectable. Thus, .9/30 or a 0.03% chance of having cancer at a yearly visit.
Oftentimes patients come in to the office with these "positive" screening results, and then when you try to find the actual tumor you can't! Since so much of cancer treatment still starts with surgical excision you can imagine the nightmare that ensues. What are you supposed to do? Give them first line chemo anyway? Wait 6 mos and image them again? It's a disaster, and I'm not being hyperbolic.
MCEDs are nice in theory, but in practice have a long way to go. The WSJ piece is just another example of low-key advertisement these companies employ to get customers (Grail is DTC light - that is, under the guise of "finding an 'early adopter' doctor willing to order the test b/c your old-school PCP wont", they take your money and then send you back to your old-school PCP to pick up the pieces).
NB also that Illumina has been ordered to divest itself of Grail b/c of anticompetition and have run afoul of both European and US authorities (https://investor.illumina.com/news/press-release-details/2023/Illumina-Announces-Decision-to-Divest-GRAIL/default.aspx), so now Grail has to prove to Wall Street that it has a customer base, that its test is so amazing that it will be adopted, etc. Hence the WSJ article. They have a fabulous PR person. They should give him/her a raise.
Cut and paste from behind a paywall:
Illumina to Divest Grail After US Appeals Court Sends Case Back to FTC
Dec 18, 2023 | Andrew P. Han
Save for later
NEW YORK – Illumina said on Sunday that it plans to divest Grail by June 2024 in either a third-party sale or a capital markets transaction.
The announcement comes after the US Court of Appeals for the Fifth Circuit issued a decision on Friday that sidestepped Illumina's grandest arguments about why the US Federal Trade Commission erred in April when it ordered Illumina to sell off the multi-cancer early detection test company it bought in 2021 for approximately $8 billion.
Illumina had suggested that the FTC's actions were unconstitutional on multiple levels. However, a three-judge panel from the appellate court said that "these constitutional challenges to the FTC's authority are foreclosed by binding Supreme Court precedent." Moreover, the decision, penned by Judge Edith Clement, said that the FTC "carried its initial burden of showing that the Illumina-Grail merger is likely to substantially lessen competition in that market […] and Illumina had not identified cognizable efficiencies to rebut the anticompetitive effects of the merger."
The court did, however, rule that the FTC had erred in its procedure and vacated the order and sent the case back to the agency for reconsideration.
Still, Illumina said in a statement that after reviewing the order, it "has elected not to pursue further appeals of the Fifth Circuit's decision."
"We are committed to an expeditious divestiture of Grail in a manner that allows its technology to continue benefiting patients," Illumina CEO Jacob Thaysen said in a statement. "The management team and I continue to focus on our core business and supporting our customers. I am confident in Illumina's opportunities and our long-term success."
The announcement signals the beginning of the end of a yearslong saga in which Illumina has run afoul of regulators on both sides of the Atlantic and experienced a shake-up at the highest levels of the company.
Illumina announced in 2020 its intentions to acquire the entirety of Grail, a company it spun off in 2016. The FTC and its European counterparts both investigated the deal's allegedly anticompetitive effects, but despite that regulatory scrutiny, Illumina forged ahead and closed the acquisition in August 2021.
Later that month, the FTC's administrative law trial began, with the agency making the case that Illumina could — and would — disadvantage other early cancer detection test makers, who relied on sequencing technology that only Illumina could provide. While the administrative law judge overseeing the case rejected the FTC's arguments in September 2022, the commission overturned that finding and ordered Illumina to unwind the deal.
Separately, Illumina received hundreds of millions of dollars in fines from European regulators for gun-jumping and in October received a second order to divest Grail.
In Monday morning trading on the Nasdaq, shares of Illumina are flat at $127.62 following a brief spike in price. Over the last five days of trading, shares of Illumina are up about 10 percent.
Really appreciate you covering this topic as I work in this industry. This paper from the Pathfinder study is what I would call putting Galleri truly in the field as a screening assay, ie phase IV. These assays never perform as well when using “known” samples. In this paper they even stratify by age as you would expect detected cancer cases to go up. But you can see sensitivity is even lower…
https://medicine.st-andrews.ac.uk/mackenzie/conference/
Good timing. Dr. Attia just posted an interview with Dr. Alex Aravanis about "...research and development of technologies and clinical tests utilizing the latest tools in DNA analysis and data science...Alex delves into two interconnected topics: liquid biopsies and epigenetics." I haven't listened yet, but if it's like all his other podcasts, it will be informative and detail oriented. (They both have degrees in engineering!) It might give you more "food" for thought on the topic to write about.
https://peterattiamd.com/alexaravanis/?utm_source=podcast-email&utm_medium=email&utm_campaign=240219-pod-alexaravanis&utm_content=240219-pod-alexaravanis-email-nonsubs&utm_source=Peter+Attia&utm_campaign=d13460a1d5-EMAIL_CAMPAIGN_2024_01_17_07_29_COPY_02&utm_medium=email&utm_term=0_-2f2efec31a-[LIST_EMAIL_ID]&mc_cid=d13460a1d5&mc_eid=ce4538805d
I coincidentally just listened to the podcast ready to share my thoughts in this comment section before seeing your comment. Dr. Attia and Dr. Aravanis provide the counter-arguments against Dr. Cifu's concerns about patient anxiety and overdiagnosis, however, I did not find their argument compelling. Please note I am just a medical student and still get tripped up on what makes a test effective (to my attending's frequent dismay..)
When questioned on whether using this test before we know it is effective at reducing cancer mortality or all-cause mortality (an end-point they basically say they find impossible to power to) is the right thing to do, Dr. Aravanis says the following (paraphrasing from 1hr 30mins if you want to check my work): If we have a test that can find early stage cancer, and a patient goes to an oncologist/onc surgeon with this information, those doctor with that information would treat that cancer. No one will say "oh the data isn't out yet on treating this based on this information, so I am going to let it be." So if the doctor is going to treat early cancer as long as they know it is there, why shouldn't we be finding all the early cancer we can find? Thereby, it is unethical not to just screen all the patients we can, whether or not we have run the study to determine if it is effective at reducing cancer mortality. (end paraphrase)
This seems to emphasize catching and treating cancers that are turtles and maybe some rabbits, but also finding birds just to find them as Dr. Prasad would put it. It is the optimistic conclusion that all data is good data. It is worrisome, but not surprising to think that the proponents of such a powerful tool do not understand that finding everything does not mean saving everyone, and that all patients will be happy to know about what is going on inside them, even if they can't do anything about it. As you said, Dr. Cifu, we may be harming patients physically, but more frequently mentally, in the process.
My optimism is that a test like this that looks for all cancers at once may actually be the first kind of screening tool that could clearly show all-cause mortality reduction rather than just cancer mortality reduction in a good trial. My pessimism is that they will not run the trial as they believe the mechanism over the evidence.
You make good points - your attending would be proud! :) I'll just push back gently on one thing you said, which was that not all patients would be happy to know what is going on inside them even if there isn't anything they can do about it. I think the point was that there's nothing much effective you can do when cancer is found late stage, but when you find it early, you actually CAN do something effective, and they can tell what type of cancer it is so they know how to treat it. But the main point I wanted to say is that Dr. Attia made clear early on that he is very up front with patients about the anxiety they might experience needlessly if they have a false positive, and also they should not consent unless they plan to do something further with a positive finding. Some of Attia's patients say, "No thanks, can't deal with that," and Attia is fine with that. In other words, he combines well informed consent with patient autonomy, and that would be my own preference. As someone who had a parent who died of cancer, I personally am more anxious WITHOUT screening. So I think it should be very much up to each patient based on their own personal and family history and preferences, leaving aside financial issues for now, as that's an entirely different basket of worms.
Richard Smith, the former editor-in-chief of The BMJ, is chatting with King Charles about his recent cancer diagnosis: https://www.bmj.com/content/384/bmj.q342
“Timeo Danae et dona ferentes”. I fear the Greeks even when they bring gifts- the Aeneid
I don’t fear Greeks. But I do fear the drug and device companies, even when they bring data
Really, though, someone needs to work faster on developing tests for silent killers like ovarian cancer. I've had three friends die of it. Can you blame a person for wanting SOME sensitivity, ANY sensitivity?
Extremely valuable article which hits on all the important points.
Specificity and sensitivity are not fixed properties of a test they have been shown to vary with prevalence and spectrum of the disease. In this case the population chosen to calculate the specificity, "healthy controls" should yield a specificity close to 100%. This wont be the case in practice. A similar test is being promoted in veterinary medicine for dogs. There is a lot of money in testing healthy individuals.
Not sure if you have seen this but it's worth a gander. From August 2023.
https://jamanetwork.com/journals/jamainternalmedicine/article-abstract/2808648
Findings In this systematic review and meta-analysis of 18 long-term randomized clinical trials involving 2.1 million individuals, colorectal cancer screening with sigmoidoscopy prolonged lifetime by 110 days, while fecal testing and mammography screening did not prolong life. An extension of 37 days was noted for prostate cancer screening with prostate-specific antigen testing and 107 days with lung cancer screening using computed tomography, but estimates are uncertain.
When it comes to screening, people completely lose the plot.
Screening ONLY works IF you can accurately detect a disease in the early stages AND treating that disease in the early stages alters the natural history and eventual outcome AND you don’t kill more people than you cure through false positives, unnecessary testing, etc.
The reason we have so many breast cancer survivors these days is that we are finding early tumours that probably never would have progressed to serious disease, but we are treating them all as if they would, while over-investigating like crazy! Big money maker for the mammogram machine makers, radiologists, pathologists, cancer society, etc. Clearly, from a business perspective, screening works if you can extract dollars from the patient’s wallet.
While I agree on the specific case of this test, I disagree with your background argument. I might suspect it depends on a general attitude towards how to practice medicine, but I'd rather avoid using claims no one made.
I don't understand the point that there's no evidence that early diagnosis improves outcomes.
First, the absence of evidence doesn't prove that there's no improvement. Second, it depends on how you act on these early diagnoses. You can do something stupid or even harmful to sell useless expensive stuff. Or you can perform life saving interventions, minimizing costs given the available resources. Early diagnosis doesn't seem to be the problem, unless you can prove that early diagnosis data is totally useless cost-benefit wise, for anything health related.
Then I don't understand the problem of overtreatment. Why an "honest" doctor would push overtreatment? The problem isn't diagnostics: it's "dishonesty" - however oversimplified is, using this concept to describe social phenomena that manifest in this context.
Not sure if you have seen this but it's worth a gander. From August 2023.
https://jamanetwork.com/journals/jamainternalmedicine/article-abstract/2808648
Findings In this systematic review and meta-analysis of 18 long-term randomized clinical trials involving 2.1 million individuals, colorectal cancer screening with sigmoidoscopy prolonged lifetime by 110 days, while fecal testing and mammography screening did not prolong life. An extension of 37 days was noted for prostate cancer screening with prostate-specific antigen testing and 107 days with lung cancer screening using computed tomography, but estimates are uncertain.
It doesn't surprise me and at the same time it doesn't sound like a real counter to my argument. Actually a study can only prove that a specific early diagnosis is useless conditioned to a specific medical follow up.
Let's abstract from a specific test and a specific diagnosis. I have an information: is it valuable? Is it useful towards any health related goal?
The answer is rarely "NO". Usually the answer is "it depends on the cost, the available intervention options, etc". A study (even a good meta-analysis) that shows that having an information didn't improve health outcomes (even the most important ones), solves only partly the medical conundrum. It just shows that either the information is useless OR we don't know what to do with it (yet).
That is to say: the study doesn't tell us if early diagnosis is indeed useless or not... Though it counters the marketing claims about the "usefulness" of specific examples of early diagnoses.
Ah marketing claims.
Most patients believe doctors know best and follow their advice without question.
There is considerable empirical research, both published and a lot unpublished (in the hands of diagnostic companies) that show the absurdity, and later futility, of using "normal" patients in the identification and validation of diagnostic tests.
The simple way of describing this is that comparing a normal control to a patient with a condition is a comparatively easy comparison, rather than comparing patients for whom the differential includes the condition, but may be other conditions in the same organ, etc.
So a priori, a study that does prospectively take all comers does not represent real world medicine and data has shown will likely over estimate the diagnostic capabilities of the test. This will be the result for Galleri.
And as a corollary, if one wants to test everybody, not just those with conditions in the differential, the test will most likely just place patients into the differential rather than identify those with cancer.
The problem I have with any cancer detection is that if the test is positive, is it really trustworthy? Next, if it is then I am faced with radiation, chemo, surgery and endless drugs with not much hope of long term survival while wrecking my body further. No thanks. There are too many natural remedies to try that will at least be as effective without incurring more damage. Ask your oncologist about them and he will give you the "baffled' look. So reassuring.
Read the book How We Do Harm by Brawley
With the current passion for surrogates, I will not be surprised if that becomes the standard of care for cancer screening one day with doctors being penalized for not ordering it. I’m also afraid of early cancer detection.
Thank you for the article. Could someone please explain to me how the pretest probability was calculated? Where did the 0.9% or 30 come from? Not doubting it, but I just didn't follow it. Thank you in advance.
Lifetime risk of pancreatic cancer about 1%. 90% of that is "regular" cancer in people over about 55. 30 is the # of years that someone is at risk of having new, treatable, cancer that is screen detectable. Thus, .9/30 or a 0.03% chance of having cancer at a yearly visit.
Oftentimes patients come in to the office with these "positive" screening results, and then when you try to find the actual tumor you can't! Since so much of cancer treatment still starts with surgical excision you can imagine the nightmare that ensues. What are you supposed to do? Give them first line chemo anyway? Wait 6 mos and image them again? It's a disaster, and I'm not being hyperbolic.
MCEDs are nice in theory, but in practice have a long way to go. The WSJ piece is just another example of low-key advertisement these companies employ to get customers (Grail is DTC light - that is, under the guise of "finding an 'early adopter' doctor willing to order the test b/c your old-school PCP wont", they take your money and then send you back to your old-school PCP to pick up the pieces).
NB also that Illumina has been ordered to divest itself of Grail b/c of anticompetition and have run afoul of both European and US authorities (https://investor.illumina.com/news/press-release-details/2023/Illumina-Announces-Decision-to-Divest-GRAIL/default.aspx), so now Grail has to prove to Wall Street that it has a customer base, that its test is so amazing that it will be adopted, etc. Hence the WSJ article. They have a fabulous PR person. They should give him/her a raise.
Cut and paste from behind a paywall:
Illumina to Divest Grail After US Appeals Court Sends Case Back to FTC
Dec 18, 2023 | Andrew P. Han
Save for later
NEW YORK – Illumina said on Sunday that it plans to divest Grail by June 2024 in either a third-party sale or a capital markets transaction.
The announcement comes after the US Court of Appeals for the Fifth Circuit issued a decision on Friday that sidestepped Illumina's grandest arguments about why the US Federal Trade Commission erred in April when it ordered Illumina to sell off the multi-cancer early detection test company it bought in 2021 for approximately $8 billion.
Illumina had suggested that the FTC's actions were unconstitutional on multiple levels. However, a three-judge panel from the appellate court said that "these constitutional challenges to the FTC's authority are foreclosed by binding Supreme Court precedent." Moreover, the decision, penned by Judge Edith Clement, said that the FTC "carried its initial burden of showing that the Illumina-Grail merger is likely to substantially lessen competition in that market […] and Illumina had not identified cognizable efficiencies to rebut the anticompetitive effects of the merger."
The court did, however, rule that the FTC had erred in its procedure and vacated the order and sent the case back to the agency for reconsideration.
Still, Illumina said in a statement that after reviewing the order, it "has elected not to pursue further appeals of the Fifth Circuit's decision."
"We are committed to an expeditious divestiture of Grail in a manner that allows its technology to continue benefiting patients," Illumina CEO Jacob Thaysen said in a statement. "The management team and I continue to focus on our core business and supporting our customers. I am confident in Illumina's opportunities and our long-term success."
The announcement signals the beginning of the end of a yearslong saga in which Illumina has run afoul of regulators on both sides of the Atlantic and experienced a shake-up at the highest levels of the company.
Illumina announced in 2020 its intentions to acquire the entirety of Grail, a company it spun off in 2016. The FTC and its European counterparts both investigated the deal's allegedly anticompetitive effects, but despite that regulatory scrutiny, Illumina forged ahead and closed the acquisition in August 2021.
Later that month, the FTC's administrative law trial began, with the agency making the case that Illumina could — and would — disadvantage other early cancer detection test makers, who relied on sequencing technology that only Illumina could provide. While the administrative law judge overseeing the case rejected the FTC's arguments in September 2022, the commission overturned that finding and ordered Illumina to unwind the deal.
Separately, Illumina received hundreds of millions of dollars in fines from European regulators for gun-jumping and in October received a second order to divest Grail.
In Monday morning trading on the Nasdaq, shares of Illumina are flat at $127.62 following a brief spike in price. Over the last five days of trading, shares of Illumina are up about 10 percent.
Really appreciate you covering this topic as I work in this industry. This paper from the Pathfinder study is what I would call putting Galleri truly in the field as a screening assay, ie phase IV. These assays never perform as well when using “known” samples. In this paper they even stratify by age as you would expect detected cancer cases to go up. But you can see sensitivity is even lower…
https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(23)01700-2/abstract