[mvadg]

Originally Posted by nagr22

My wife came across a homework problem in her statistics class

You are missing the forest for the trees. Simple fact - more tests mean higher accuracy, everything else remaining the same.

[nagr22]

Originally Posted by mvadg

Simple fact - more tests mean higher accuracy.

I disagree. Accuracy of a test is only dependent on the technology behind the test and the skill of the tester. Until the disease reaches community spread stage (god forbid!), more random tests equals more false positives and that's the scientific theory. The keyword here is random tests. We definitely must trace and test all possible suspects who have come in contact with positive cases and those with travel history. But we should not randomly test everyone for the disease.

I am not a doctor so I can't say if this theory is right or wrong (It makes all sense to me on paper). That's why I requested the opinion of doctors who know these aspects practically.

[Behemoth]

At least some good news: In a study shared by Science direct, Ivermectin (which is an existing anti parasitic drug) has shown promising results against the Covid 19 Virus. If the labs and doctors were to conduct further testing on this and it could be further proven to be effective, how soon would it possible to get this released for treatment ?(Question to Doctors)
https://www.sciencedirect.com/scienc...66354220302011

For other people interested, here is a Scientific paper from International journal of Molecular sciences which outlines the Covid19 Virus and also the current prospectives which are undergoing Clinical trials.
[ATTACH=ijms-21-02657-v2.pdf]1994202[/ATTACH]

[vibbs]

Originally Posted by nagr22

I disagree. Accuracy of a test is only dependent on the technology behind the test and the skill of the tester. Until the disease reaches community spread stage (god forbid!), more random tests equals more false positives and that's the scientific theory. The keyword here is random tests. We definitely must trace and test all possible suspects who have come in contact with positive cases and those with travel history. But we should not randomly test everyone for the disease.
.

I don't understand the logic. You say accuracy of a test is dependent on the technology and then say more random tests will give more false positives. Is this not the case if only the technology is flawed?


What about confirmed true positives not being registered due to lack of random testing?

You are assuming there is no community spread yet. But the only way to confirm this is through random testing!

[skdking]

Originally Posted by nagr22

My wife came across a homework problem in her statistics class and we were fascinated by its relevance to the current situation! ....................symptoms? Are they false positives???

Dear nagr22,

A fresh dose of sanity in the cacophony of 'testing, testing, testing'.

However we should continue to discuss all options, here.

[balenoed_]

Originally Posted by nagr22

I disagree. Accuracy of a test is only dependent on the technology behind the test and the skill of the tester.

I do not think he meant the accuracy of each tests. I think he meant to say the accuracy of the number of cases. The more you test, the more you know where you stand.

[DCEite]

Originally Posted by Behemoth

At least some good news: In a study shared by Science direct, Ivermectin (which is an existing anti parasitic drug) has shown promising results against the Covid 19 Virus. If the labs and doctors were to conduct further testing on this and it could be further proven to be effective, how soon would it possible to get this released for treatment ?(Question to Doctors)
https://www.sciencedirect.com/scienc...66354220302011

For other people interested, here is a Scientific paper from International journal of Molecular sciences which outlines the Covid19 Virus and also the current prospectives which are undergoing Clinical trials.
[ATTACH=ijms-21-02657-v2.pdf]1994202[/ATTACH]

The problem with the study , as many "rationalists" would say, is that it is "in vitro".

Even then, i think there is something else going on here. Lots of medicines like Hydroxychloroquine, Remdesivir etc have been found effective in small trials or in-vitro studies.

But none of them are making it in the "official" list of CoVID-19 treatment or even large clinical trials. Not sure why. It makes me wonder, whether the Vaccine lobby does not want a $5 pill to do what a $50 vaccine would potentially achieve in future.

[mvadg]

Originally Posted by vivek95

{I took the liberty of reorganizing your post for my own comprehension, hope you don't mind.}
{ Problems}

1. The pathetic doctor to patient/ nurse to patient ratio
2. overall inadequate healthcare infrastructure
3. Violence towards Doctors to the extent of killing them
4. absolutely no support or protection from the Government.
5. You may also start reading news of people dying on roads out of hypoglycemia

{Solutions}

1. Practical solutions could be Government providing PPE kits in abundance
2. rapid Covid-19 testing (free of cost),
3. creating larger isolation facilities and thereby
4. ensuring that all hospitals big or small keep functioning w.r.t their Emergency departments and ICUs.
5. I am not talking about ventilators because that's a distant dream.
6. Bigger well equipped hospitals should reserve themselves for Covid patients/all other critical conditions.
7. moderate kind of ailments yet that are dangerous if left unattended on time.
8. Dialysis and chemo centres be rather open, please. Government is doing its best

but we are running out of time.
--Dr. Vivek

Firstly, thank you for your dedication and sacrifices, we need people like you in every field! As an earnest young doctor you've put forward your first hand experience from the frontlines of this fight against the SARS-CoV2 and it merits serious attention from those at the very top

• - From what I can see, in the Problems section there are easy solutions to some of the problems, others need long term plans - the government can definitely provide medical staff and hospitals with 100 percent protection with armed military personnel for both the medical infrastructure and each and every staff member.
• - In the Solutions section - I am yet to understand why PPE is not freely available, months after OFFICIAL recognition of the problem.
• - I think most of us can afford to pay for tests
• - Correct me if I'm wrong, but is there any need for medical personnel to supervise patients who cannot be provided ventilators or other complicated interventions? Can the vast majority not simply be isolated and provided palliative care by their loved ones - that too away from hospitals?
• - I think every patient and doctor/staff entering a non-SARS-CoV2 area of any hospital must be first tested for SARS-Cov2 and placed under strict quarantine for 21 days (regardless of the test results) before being admitted for any kind of procedure at a hospital to safeguard hospital staff. This will keep the hospitals in business and running - there is no way we can shut down these essential services!
• - Doctors attending to SARS-COV2 cases must be provided maximum facilities, protection and isolation - we should consider them top-notch, elite soldiers and accorded the same respect and stature. In fact, I am disappointed you are venting your frustrations here when you and your colleagues should have a direct line to the PM - I'm sure there are no more than a few hundreds of you and he should set up a time to review your inputs on a daily basis.
• - Even before the SARS-CoV2, I've seen my cousin get infected with TB as an intern, something I felt was inexcusable given the kind of knowledge and technology we have these days. Clean rooms should not be reserved only for silicon fabs!

>>> ****but we are running out of time*** - Absolutely!
TBHPians please feel free to poke holes in my post!

[mvadg]

Originally Posted by nagr22

...

I think you read my post before I edited it, I added "everything else remaining the same". Sorry, if I confused you. I just used basic statistics.

[joslicx]

There is a huge risk of hospitals and test centers themselves inadvertently spreading the virus to healthy people. World over thousands of doctors and health care professionals have been infected. They are the riskiest places on earth.

I wouldn't go out for any testing right now, not until I was absolutely displaying symptoms.

[nagr22]

Originally Posted by vibbs

I don't understand the logic. You say accuracy of a test is dependent on the technology and then say more random tests will give more false positives. Is this not the case if only the technology is flawed?

Bayesian theory is very counter intuitive and it does sound illogical until we understand it. Try reading the link given in my original post again and you may understand better. Let me try to explain m with another example which we can all relate to.
Assume 0.1% of population is pregnant at any given point of time. Assume that the best pregnency test kit available in the market has an accuracy of 99%. This means, if the test says positive we can be only 99% confident that it is positive. Similarly if the test says negative, it is again only 99% confident. 1 out of 100 times the test result is incorrect (false positive or false negative, could be anything). 99% is a very accurate test right? Nope and you'll know why.

Let us say we take sample of a person at random and test for pregnancy (assume we don't know any history, not even whether the sample is from a male or female).

If the test is positive, it DOES NOT mean that there is 99% chance that the person is pregnant. This is because we already know that there is only 0.1% chance that a random person is pregnant. The fact that the test came positive is an additional information. Using this information, if we calculate from Bayes theorem, the probability that the person is indeed pregnant is given by the formula,
= 0.99*0.001/(0.99*0.001 + 0.999*0.01)
= 9%

So a random pregnancy test only gives 9% confidence. If we repeat this test for the same person, then the confidence level goes upto 91%! The same is true for a negative result as well. We have to run the test on the same person twice in order to successfully tell whether he/she really is positive or not.

Suppose we have additional information about the sample. Say we know it is from a female. Which means now the probability that a random sample is positive becomes 0.2% (Assuming 50:50 male:female ration). Lets say we also know that the said female is of age 20-30 and is married etc etc. The more history we have, more is the confidence level.

Please read this article for more details about formula and explanation

The example applies to covid as well and this is why they say a person has to test negative twice before he can be discharged from isolation ward. My point is that more random will lead to more inconclusive (false positives and false negatives) results. I strongly advocate for targeted testing which is what is being done currently (at least on paper).

After understanding Bayes theorem, My faith in the medical committee who are managing covid test protocol has been multiplied manifold! It feels reassuring to know that they are calling the shots based on scientific research methods.
My only hope is that those who are implementing these methods are doing a thorough contact tracing, checking travel history etc. so that the results are accurate. Another question is whether the suspected contacts are honestly revealing their histories.

[JithinR]

Delhi Pizza delivery agent tests positive for COVID-19.

https://www.ndtv.com/india-news/coro...home-topscroll

[volkman10]

Kerala flattens the coronavirus curve but must remain vigilant as 'import' cases still dominate - A case of concern at large.

Kerala had the most number of COVID-19 cases 20 days back but reported only one new positive case on Wednesday. With high testing and low fatality rates, Kerala’s numbers show promise but some concerns remain. Recent cases are still “imported” from outside India, even though international travel halted on March 22.

Considering international flights were stopped on March 22 and quarantines were enforced, these delayed confirmations are a cause for concern. Countries like Japan and SIngapore which successfully flattened the curve initially are also seeing a second wave of infections.

Link

[Lobogris]

Originally Posted by joslicx

I wouldn't go out for any testing right now, not until I was absolutely displaying symptoms.

Even with symptoms, what would you gain from the test? Unless you are seriously sick to the point where you can’t breathe or have extreme fever, you are better off just self isolating yourself as there is no medicine or cure available in any case. I don’t see the logic in hospitalization of so many people needlessly. It takes space away from people who have other urgent illnesses and it increases the risk of infection for everyone from doctors to testing staff, nurses, ambulance staff, other patients and cleaning staff. For example if Bombay has 2000 infected and we have hospitalized all of them when only say 100 are in a serious condition then we are wasting 1900 beds and needlessly exposing so many additional staff required to look after these extra 1900 people. Sweden has such a policy where they discourage tests or hospital visits even if you suspect you are infected. Of course if you are poor and don’t have a separate room available then that is a different matter.

[nagr22]

Originally Posted by vibbs

I don't understand the logic. You say accuracy of a test is dependent on the technology and then say more random tests will give more false positives. Is this not the case if only the technology is flawed?

To explain in even simpler terms, if a test is marked as 99% accurate it means that if I give 100 samples which are already known to be positive, then the test result will show positive 99 times and 1 test will be a false negative. This is due to the way they test in the laboratory.
Now how accurate is the test result when I know that only 1 out of the 100 samples is actually positive? This seems like an easy question but it's not. Bayesian statistics is a subject designed around problems like these.

Now how do we know what percentage of the community has Covid? We can't know with confidence unless we test the whole country, twice! Can we do 2.6 Billion tests? If we do and find out that many people have it, what's the course of action? We will most likely ask them to self isolate, which is what we are doing now albeit for everyone irrespective of positive or not. Can we lockdown forever? Nope. We do our best to trace and isolate all the positive cases and pray that we did it right.