Milestone update - 100000 covid test at 2100 avg rates + antibody tests 25000 carried out so far( combined with previous post that’s 15K test in 6 days as additional I.e. 2.5K per day run rate for Antibody test)
3 Likes
@vivek_mashrani and folks who’ve invested, how do you look at threats from the likes of https://athelas.com/ ?
Deep learning is going to make a lot of diagnostics much cheaper, to the point where you might not require a diagnostic chain. All you will need is a healthcare professional to explain the report and even that would most likely be automated… The only reason not for automating would be that humans arent too comfortable talking with bots.
What are your views on this from a long term perspective?
This might not be a threat in the short term, but as soon as these tests are proven reliable (which they will be for sure), many diagnostic tests are going to disappear very very quickly
The problem with disruption is that we usually get it wrong in both ways - we tend to overestimate it or underestimate it.
For example, if you do a Google search on article from 2009, you will see that most experts thought our roads would be choked with Tata Nano and the autorickshaws would become extinct. There was even a book written about the Nano - Nanovation, if I am right. But as of today, the Nano is extinct and the autos are still around.
On the other hand, even Steve Ballmer underestimated what the iPhone can do. However, it over exceeded even the wildest imagination. Think of how many businesses have come up that are built around the smart phone: Uber, Ola, WhatsApp, Swiggy etc.
About 2 years ago, Outlook Business covered an article about a healthcare startup- SigTuple, which was like a successful Indian version of Theranos. You can read that article here.
And about a month back, The Ken also wrote about SigTuple saying how the Covid has dealt almost a death blow for it. You can read it here. The startups are having existential crisis but the established players aren’t.
Disruption is undeniable. When will it happen, in what form and shape and what company is hard for anybody to predict or model.
For now and for the limited period that we can see, we feel that the diagnostics market is a huge pond and many fishes can coexist and grow at fast rates for a significant period of time.
13 Likes
Agreed but disruption leads to binary outcomes most often.
Nano got screwed because of what happened in Bengal and shifting the factories to gujarat where they didnt get the incentives they were getting in bengal and had to recover the costs that they bore for setting up the factory and then dismantling before selling even a single car. This meant huge delays in delivering the cars, increasing price to over 1 lac and on basically they couldnt fulfill their promise of “lakhtakiya”
This case is different. If there’s a machine which costs a few lacs and can do the same tests which originally were done by really expensive machines with similar accuracy levels (perhaps slightly lower accuracy level will also be acceptted), then any small diagnostic lab should logically go out to buy this cheap variant. Faster results, higher margin, no cash outgo to B2B players.
As said earlier, disruptions have binary outcomes and a B2B model is even more vulnerable to this sort of disruption. It would take a few quarters to affect the topline of B2C players because customers would probably want to stick with reputed players but this is also temporary. Binary outcomes might be acceptable with scandal investing and grave dancing but if you’re paying prices which are reflecting quality, then binary outcomes are perhaps less preferable.
All of this is under the premise that this field is actually going to get disrupted in the medium term and as Prof Bakshi says, always try to figure out the worst possible scenario and see if that’s acceptable. In this case, worst possible scenario is unnerving imo. The fact that there are already startups that are catching the limelight in India just accentuates this feeling.
Disc: not invested but spent the past few days studying the industry & business. Also, am a ML enthusiast so perhaps biased
Point of care devices have been discussed in this thread in the past - please go through the entire thread. Not interested in repeating the blurb
Few lakh rupees for that flash device which will give only white blood count and neutrophil count!!! Wow…good luck
For Rs 300-500 one could get full blood count analysed - it will give you the following reports haemoglobin, MCV, RBC count, white cell count, neutrophils, basophils, eosinophils, platelets, blast cells, reticulocytes. Much more informative than the white cell count and neutrophils only from that flash device…
What about other routine blood tests - glucose, liver test, kidney test, vitamins, lipid, iron and other tests. That device can’t test these…
You’re right those devices cant test blood tests - glucose, liver test etc…yet.
The intention of the post was not to malign the present strength of the company but rather point out the threats that are coming its way. Having followed the machine learning space, just wanted to make the community aware of possible disruption.
Biotech and Deep Learning for the healthcare space is developing rapidly.
https://www.nature.com/articles/s41598-018-34300-2
This space is developing faster than most people believe and understand and smart devices such as smart watches are providing companies with alternative ways of measuring and monitoring health of the users. The way one walks can predict how likely they are of having dementia. A US based study has found sleep data and heart rate from fitbit are better at predicting flu more accurately than current methods. There are so many more examples.
This space is poised for disruption and one would be wise to understand the threat of disease diagnosis from these relatively unorthodox sources.
As they say, the enemy you know is not the real enemy. The unknown is where the real enemy lurks
AI is being increasingly utilised in Radiology, but not yet in blood testing. That is an important point to remember. Thyrocare is a blood testing diagnostic company (nueclear is only <10%)…
In radiology too, AI is used to answer specific questions using certain/specific radiology tests (is there breast cancer, is there lung cancer etc) - AI is not so useful in day-to-day routine radiology tests. AI is definitely making the job of radiologists easier - pattern recognition alerts the radiologist of specific abnormalities. The claim that AI will make radiology jobs redundant is greatly exaggerated…
For eg - if I go to a Dr with severe cough and fever -chest x-ray will have to be done. X-ray is still the best screening test (pneumonia, heart failure etc) - AI is not useful here. Blood tests also will have to be done - full blood count is informative to differentiate bacterial from viral infection, biochemistry tests will give information about electrolytes and other vital organ function (Liver/kidney tests are sometimes impaired in severe infection). Blood tests will also tell if I am a diabetic - sometimes diabetes and other disorders are diagnosed incidentally when we do testing for other reasons. Blood tests will also reveal whether I am iron or b12 deficient. AI will play no role in the above scenario. Point of care (PoC) devices are absolutely useless in the above scenario… one PoC device (expensive worth a few lakhs) will give out white blood and neutrophil count and another PoC device will give out blood glucose tests (only a few thousand rupees I think) - for all the other blood tests one has to rely on the routine means of testing - what will you do if you have to make this decision?
AI is very useful in answering specific questions using certain special radiology tests - for eg breast cancer screening (this is done in females 45-75 years of age - not routine practice in India). Mammograms are sometimes tricky to interpret in females with dense breasts - AI is very useful and improves the correct diagnosis greatly.
Thyrocare shareholders hold on to your shares. AI is not going to disrupt blood testing any time soon. Thyrocare is not into radiology testing (except pet ct scan business)…
Discl - very biased
3 Likes
There are several positive/price sensitive details in this announcement and this was released on Friday 10 July after market hours!! My suggestion to the fellow (or aspiring) shareholders is to read the announcement in full (including the last page) and make sense of the information in that document.
No wonder the stock price has appreciated by 15% in the last 2 days. Latest shareholder statistics indicate that retail holding is the lowest ever at 5.91% which is a shame…
Discl - hold and have been adding since IPO
Agreed to all your points but the thing is, AI is changing how diagnosis is done.
Photographs of the eye accurately predict 93% of the time if a person smokes or not. How many of us would have thought this possible just 2 years back? (https://www.nature.com/articles/s41598-019-43670-0)
“artificial intelligence (AI) screening system is capable of accurately discovering diabetic retinopathy 95.5 per cent of the times.” all it needs is photograph of the eye
(https://economictimes.indiatimes.com/magazines/panache/medical-marvel-new-ai-system-can-accurately-detect-diabetic-retinopathy/articleshow/71615821.cms)
will this test be cheaper than existing tests? You bet. Will AI find ways to figure out cheaper alternatives to existing tests? Probably. Will they too be cheaper than existing tests? You bet. Will it be good for any company in this industry? Probably not
Having spent the past few days studying the company and the promoter, would definitely have loved to go in on this. The promoter is an absolute genius, intelligent fanatic… But the threat is real. Maybe not for the short-medium term but that’s precisely the problem, we dont know exactly when things will go south.
1 Like
Are you able to explain which diagnostic tests are under threat using the examples you are citing…Let me try to explain
Photographs of the eye accurately predict 93% of the time if a person smokes or not. How many of us would have thought this possible just 2 years back? (https://www.nature.com/articles/s41598-019-43670-0 )
*he above example of yours - are you able to tell me which blood tests are under threat? I will tell you - none. Why do you need to take a photo of the eye to tell whether he/she is a smoker or not - you just ask the person about it. Photo of the retina is not required and a wasteful test for this indication
Your next example…
“rtificial intelligence (AI) screening system is capable of accurately discovering diabetic retinopathy 95.5 per cent of the times.” *all it needs is photograph of the eye
(https://economictimes.indiatimes.com/magazines/panache/medical-marvel-new-ai-system-can-accurately-detect-diabetic-retinopathy/articleshow/71615821.cms)
*Please tell me which diagnostic tests are under threat?? I will tell you - none. Blood tests are not used to diagnose diabetic retinopathy!!
AI screening system is great in early diagnosis of retinopathy - that is great. Who benefits? Patient - of course, early diagnosis will save his vision for longer. Who else will benefit - Ophthalmologist!! Why - once a person is diagnosed with diabetic retinopathy annual monitoring is required by an ophthalmologist so that he could intervene in case of advanced disease or complications. What is the point in diagnosing early if complications of retinopathy are not managed properly - you can’t avoid seeing ophthalmologist in this scenario (in fact ophthalmologists also benefits because of earlier diagnosis!!)
In both the above scenarios, diagnostic companies are not affected in any way!! Also in the second scenario, diagnostic companies might actually benefit since diabetes will have to be controlled more aggressively and that means more frequent blood testing for glucose control and to monitor other end-organ-damage
Please give us proper examples of how blood diagnostic industry is going to be adversely affected by the advancements in AI or ML (machine learning)…
2 Likes
the reason for the company/industry to exist is to provide an update on the health of the person.
The point of the examples was to show that indirect and significantly cheaper means of solving this purpose exist already which completely bypass the blood diagnostic industry and are non intrusive. (Horse carts exist even today but as a novelty. Cars fulfilled the purpose of the carts, and they were cheaper and more reliable)
https://www.nature.com/articles/s41598-019-56927-5
Detecting Hypoglycemic by monitoring heart rates, no blood involved.
https://www.nature.com/articles/s41551-019-0487-z
Detecting anemia using retinal images.
Fundus cameras are available in every eye testing centre, do these eye testing centres become competition for conducting the anemia test? The cameras cost 50-60,000 rupees btw
Cardiovascular disorders are monitored by many smart watches
Ultrasounds to predict thyroid cancer, with true positive detection rate of 97%. No blood involved.
And even if these examples arent adequate, it is scary to see so many research papers being published in this field, and most of these have been published in the past 2 years. Does that not show how close people are to a major breakthrough?
1 Like
Seems the point of discussion is interruption in business & certainly it would be there but big question is WHEN?
& even when it happens … first level diagnostic will get split in above versions & second level (detailed investigation) remains with current.
Now next question will require data, how many first level diagnosed patients require second level of detailed diagnosis. If someone has, kindly share.
Overall AI / ML can interrupt entry level or first step of diagnosis business but its going to take longer & its always good to keep close eye on that… mainly companies investing on R&D on similar lines. If we see any business model doing wonders, we must certainly catch them as investors… money making is important
1 Like
This discussion is not helping in any way in terms of understanding the company or the sector.
As for ML and AI it has a lot of potential but till the time a commercial challenger comes against thyrocare, we can dicuss it at that time, till then it is all speculation.
2 Likes
I agree that discussions got derailed and we have gained nothing - I wish posters put some thought into the contents before publishing here.
Back to the topic - after disruptions in Q4/Q1 this year covid is actually going to contribute significantly to the topline and bottomline. Covid nasal swab PCR daily testing capacity in India is 5000 per day and in Gulf is 5000 per day (total 10,000 per day)!! This is as per their announcement.
Blood antibody testing also will rise exponentially since there is no rate limiting step for blood tests.
All the other routine blood testing will continue to be adversely affected due to lockdown situation across the country.
1 Like
This is bad publicity for all lab testing companies. It is seriously undermining credibility of these labs…
Please read this announcement and you will have the answer…
3 Likes
Worth listening to Velumani speaking about the future of his business and how he is increasing the employee strength
2 Likes
Well done to all the shareholder who kept faith!
Result will be announced on 31 July. Just trying to guesstimate net profit
Assumption - 90,000 covid pcr tests for the quarter (profit per sample Rs 1000) and 30,000 blood antibody tests (profit per sample Rs 100).
9-10 crores net profit (ignoring impairment of Nueclear business if any) and also assuming that there was no non-covid blood testing. I know, 9-10 crores doesn’t sound like a big number (in fact 1/3rd of their usual net profit) but they are now doing 10,000 pcr tests per day and I presume the capacity is increasing every day…
Discl - very biased
2 Likes
Just for your information…Dr A.Velumani have declared that his company would not make any profits from covid tests done for next 1 year. Hopefully, he would be profiteering from Antigen test being done from last three weeks.