That doesn’t explain them having data on 94% of US patients or not having 4 Australian hospitals that add up to the number of cases they had from Australia. It’s fraud, pure and simple. Some enterprising lawyer representing the family of a patient that died after his physician refused to prescribe HCQ based on having read the Lancet paper should sue Surgisphere, the authors and the Lancet. Civil suits are entitled to discovery. If it doesn’t end up in the court of a bought and paid for judge it would make a great trial.
In present times, models and polls are like rear ends. Everbody has one, they all stink, and the “media” savors the acrid stench. Money talks, truth and trust walks.
Its really easy to look at the people who do the wrong things, illegal , unethical and otherwise. However, if you had the power, what would you do? its easy to think , “Oh, Id do the right thing for sure.” I am one of those, that nothing will make do otherwise, but , I think i am far from majority on this.
So the question is posed, Lets say you are tedros, and the powers that be come to you and say look at all this data , “we have to stop the HCQ treatment - its really harmful - we are being reckless. There is no real data. We need to tell people…” ( tedros - "Maybe we should do some official studies? ) “No, we don’t need that. We have evidence that this “drug” will do a better job and we can make a lot of money… I will get you some… !” What do you do? Tedros - “Well maybe it(HCQ) is reckless. All these people don’t need some repurposed drugs - we really dont have good studies; it could be dangerous… sure… ok… lets do some research on the new (money) stuff.” And find something new for a new problem. ( maybe you try to convince yourself what is right ) - what is right by other people in position of influence about your job , your accolades… Guess what, no problem, “I will play ball - I cant fight them all anyway, maybe they know better, might as well buy my kid a BMW and walk him into the best universities… I am not really gonna worry about the small details…” ( it might be pretty simple for most to think they are doing the best or they need to play ball with others - politics 101 at the office ) - “this is how I got where I am.”
Seems its not hard for “human” psychology to take over - and trick a person into doing really really wrong stuff. You really need to stay objective and open. I know Chris speaks about it a bit. You have to listen, you have to stay objective and remain unbiased and willing to accept something other than what you expect. You have to be able to change and be adaptable. ( but you have to have a set of principles, values, beliefs, and constitutions that guide you )
I think in your mind - it would not be a far reach to justify doing things you think are best - because others tell you they are best. Its very hard to know yourself and know you are sure , while being certain that you are acting in the interest of all. ( even when that is really your job ) Bottom line, if you are making decisions based on what others tell you and not what you know and expect, you are likely doing it wrong.
Well, I am sure some of the well-meaning do-gooders could get lost here in this logic even if they were well-intentioned.
Lets take this question one step beyond motivation of pleasing those that give you power. and the money that comes from that. But perhaps instead of reward based motivation for actions, you are threatened. Now what?/ threatened to be removed from your position? your career trashed? Or perhaps your own safety or the safety of your own family? Now what?
What would you do?
I’m guessing Surgishpere is pretty small fry- 6 employees? But sue them, and I’d wager Big Pharma would step in and have their legal teams ensure the case never came to trial- you know, rule of law, justice, and all those fantasies…Aloha, Steve.
How does AI fill in missing data? If you have a spread of data along a sloped line, what is the bias that AI would bring to calculating the “infill” points Surgisphere needs? This Lancet retrospective survey shows much tighter variation in data spread of the 23 factors than what appears in ordinary matching for data sets this small according to “Arkanicide”. Does the AI program used here accept infill points and reject outliers? Is it a standard AI program accepted as a research norm? Uh, no. It was a proprietary program created by Surgisphere and not peer reviewed by anybody!
Filling missing data with AI is a joke for real research and data analysis. You must really program that AI with a basic set of logic to build off of… It can be biased and manipulated. AI for missing data, just doesn’t work in real world studies. It is an interesting exercise in modeling… for entertainment value. However, it would have exceptional value if you actually wanted to game a study, if you were lucky enough to get stupid people to think your study( and data ) was legit.
I have been working with an AI system for almost 2 years (industrial not medical data) and I can tell you that if these people are claiming they wrote the code, trained the AI and got actionable data they’d be willing to state their literal lives upon in only a few months it is definitely time to waive the BS flag! Also, from 96000 individual cases to still over 3000 similar cases using 23 factors with those BS st dev and mean values is a total fabrication. I have a couple hundred thousand “cases” in my data and being industrial data it is designed to be similar because we are making standardized products, unlike people. I cannot put in more than a handful of factors before my data sets get trimmed down to essentially meaningless numbers of “cases.
also, it is extremely easy to introduce accidental bias and even easier when you want to do it on purpose, which is what I suspect happened in this case. I enjoy working with my AI system and it is extremely helpful in going through large amounts of information, however, all results from the system should first pass “the smell test” before you act upon them.
How does AI fill in missing data? If you have a spread of data along a sloped line, what is the bias that AI would bring to calculating the "infill" points Surgisphere needs? This Lancet retrospective survey shows much tighter variation in data spread of the 23 factors than what appears in ordinary matching for data sets this small according to "Arkanicide". Does the AI program used here accept infill points and reject outliers? Is it a standard AI program accepted as a research norm? Uh, no. It was a proprietary program created by Surgisphere and not peer reviewed by anybody!Let's go one step further. How many people on Surgisphere's 6 person squad are identified on LinkedIn as having any big data or AI background? Zero. You've got the main MD Desai, then a business development dude, a couple of "science editors" which I presume means "marketing copywriters," a VP or marketing & sales, and a general mgr. That's the team. So who actually wrote all this proprietary AI code? Desai the cardiologist? I'm guessing it was outsourced to India or something. Since this is a poorly funded start-up it was probably done on the cheap. You know why they are resisting every effort to peer in their black box? Probably because it is mostly an empty void. Heck, might even be excel spreadsheets with some half-fancy macros. Seriously - where did this world-class AI program come from? That's a legit question.
You can only use complete patient records in your training data set. So either you have to throw out all patient records with at least one missing value or start throwing out the variables with the most missing values especially if they tend to be missing for patients for which most other variables are good. Then you have to reserve part of your training data set for testing (verification) of your model as you iterate it during training. Then you have to repeat for each variable that has missing values. And you have the problem of representativeness - if most cases of a missing value tend to be from a specific group - i.e. missing smoker/nonsmoker records are mostly from a specific continent, gender, race, age, etc. or combination of two of those, and there aren’t a lot of complete records from that group, your model will not be very good.
QB-
Agree with your analysis, especially the bit about “representative-ness.” For instance, when training up my candle models, I first used all the candle prints from all 1500 items in SPX, smallcap, and midcap as training data. Problem is, on huge moves (say, during a crash), all the candles of all the items looked identical on those big days, so I’d get (say) 1500 pretty much identical takuri line candles all with the same pattern right at the low, which would skew the training result way too much. I ended up having to bucket by day, and then pick (at most) 20 samples at random from that day. So - representativeness is exactly right.
takeaway: you have to understand your data, play with it for a while, and even then it’s a tough problem.
That brings me to 2:
Chris- There were no AI software engineers on the team.
That’s a fail too. Engineer needs to understand what the data really means, else the result won’t be right.
I’d add 3: importing 671 sets of hospital data (presumably, each in its own format) is … a quagmire. It would take a team of people months and months to get this right. Especially reducing the text fields in 10 different languages to something resembling coherence. Months and months. A team. And they have 6 people.
The more we all poke at this, the more pathetic it looks.
In the CPA world, we would call this a pass-through company. It doesn’t really exist. It’s just a circle jerk. Surgisphere is nothing more than the dr who authored this study under the guise of looking like his study is genuine while trying to escape the truth about his fraudulent data/ Basically he cannot say we got our data from blah blah… he himself is the data company… Nothing less. ( AND this should constitute a HUGE CONFLICT OF INTEREST ) ----- STUDY FAIL. ( why would a cardiologist or researcher own a data company? ) Marketing research companies… sure. Insurance companies, sure… researchers , doctors ( big no…no… )
Why would anyone use AI information to treat a patient? aren’t we dealing with someone’s life here? We are not serving search results or feeding someone an advertisement… What I feel is, this study could be defined more as a model…
I know what I would do. Your question of ‘What would you do?’ when confronted with an ethical dilemma which could affect one’s future, may apply to those who have not sworn to follow a code of ethics for their profession. But in Medicine, as in Law and Accounting (in Western society), people in these fields normally take an oath when they join the profession, to always follow the ethical guidelines set out for their profession, which is then overseen by a governing body. Part of ‘doing the right thing’ involves making oneself aware of all the information, and being open to considering new information as it arises, regarding an issue, and then weighing this against the various standards which apply to their profession.
In Medicine, Law and Accounting, there is no room for corruption, fraud or other behaviour which does not uphold the highest level of ethics. There is no justification for it, and no situation under which it is acceptable. Otherwise we do not have a civilized society. This value system doesn’t hold in all countries in the world, but it does in Western countries, and there are usually stiff penalties for unethical behaviour.
This is one reason why administrators who are in a position to set standards and influence the members of these professions, should themselves have sworn to such oaths. If they are not members of the profession and/or their current position is now purely of a political nature, then it may not be appropriate for them to be in such a position of influence.
If people agree to walk down that path where fraud, corruption, and negligence are acceptable in order to keep one’s job, be promoted, or meet some other goal, then our society turns into one that resembles some countries on the other side of the world, where nothing happens without a bribe, and this is acceptable and part of the way things work.
What if you’re a 14 yr old girl hired to massage an old rich guy who then wants more than a massage?
What if you’re a politician (a president) when three sailors are killed in a terrorist attack? What do you do?
Distract, delay, delude, play golf, go along to get along?
I got a lot of it; and all the examples just rail my meter. So all the examples of govt. halting treatments leads me to believe the treatments are both safe and effective. If easy treatments were unsafe and/or ineffective the evidence would show and fraud studies would not have to be generated. Small unheard of company generates huge study leading huge manager decisions. Air America without the drugs… Oh wait it is about drugs.
My experience with AI algorithms shows me that they work fine with mildly nonlinear problems, but when you get into heavily nonlinear ones an expert’s intuition always wins. Do you really think AI will “think up” more advanced AI algorithms? If you do, you have been watching too many Star Wars movies.
Trump’s war with Twitter’s fact checking is the latest version of book burning. “The Buck Stops Here” has become “The Buck Stops Over There in China, Iran, Venezuela, etc etc”
Yesterday Surgisphere posted a letter on their website responding to criticism of their Lancet study.
https://surgisphere.com/2020/05/29/response-to-widespread-reaction-to-recent-lancet-article-on-hydroxychloroquine/ ...In so doing, we join agencies including the U.S. Food and Drug Administration, as well as several other observational studies reported in the NEJM, JAMA and the BMJ, each of which have pointed to either no benefit of the drug regimens using hydroxychloroquine or chloroquine or even a signal of potential harm (also see this paper). The Brazilian CloroCovid-19 Study on chloroquine diphosphate was recently stopped in Brazil due to the noted safety hazards associated with this drug.Wow.. they really tipped their hand by even mentioning the Brazil study. I profiled and critiqued this study when it first came out - the high dose arm of the study, which did in fact kill patients, was so high as to be intentionally designed to fail. Others have said it was tantamount to malpractice. For a Cardiologist like Desai to reference this ultra high dose Choroquine (not hydroxychloroquine) study without pointing out the obvious fact that it is in no way representative of the way the rest of the world is using HCQ shows that he is not pursuing sincere science. He is instead, knowingly pursuing a false, anti-HCQ narrative. Here is what I wrote at the time;
https://peakprosperity.com/forum-topic/hydroxychloroquine-vs-the-globalist-deep-state/page/9/#post-558716 Dr. Zelenko dosing of hydroxychloroquine; 200mg x2 daily = 400 mg/day, usually for five days. Brazil study low dose of chloroquine phosphate; 450 mg x2 daily = 900 mg/day, five days Brazil study high dose where people died; 600 mg x2 daily = 1200 mg/day for 10 days!!!! OMG, images of the AZ fish tank cleaner death pop into our heads! Knowing what we know, that this drug has a long half life, and that a, “normal” dose of HCQ for arthritis is 400 mg/day (see below) … exactly what did the Brazil doctors think was going to happen in the high dose arm of their study. Looks like it was designed to fail, to me.
In so doing, we join agencies including the U.S. Food and Drug Administration, as well as several other observational studies reported in the NEJM, JAMA and the BMJ, each of which have pointed to either no benefit of the drug regimens using hydroxychloroquine or chloroquine or even a signal of potential harmLogical Fallacy: Appeal to Authority "Hey! We're saying the same thing that these Really Fancy Journals are saying! We think HCQ is lame too! We are so mainstream it hurts! This is how you know our study can't possibly be flawed." Now I'm just piling on.
The picture is coming clear - “Dr” Desai is a serial fraud, at best a person who plays fast and loose to achieve his aims.
Check out this background from The Scientist:
Surgisphere and its founder Desai is now focused on Surgisphere, which currently has 11 employees, he tells The Scientist. Surgisphere’s website states that, “When Dr. Sapan Desai founded Surgisphere Corporation, the mission was simple: to harness the power of data analytics and improve the lives of as many people as possible.” Desai tells The Scientist that his company has always been involved in data analytics. When Desai established the company in 2008 while a surgical resident at Duke University in Durham, North Carolina, Surgisphere Corporation’s most visible activity was marketing textbooks, produced by Surgisphere, to medical students. Sapan Desai says he can understand people’s concerns and that the burden of proof rests with Surgisphere. Reviews of the company’s products on Amazon are polarized, and a handful of positive reviews that appeared to impersonate actual physicians were removed when those doctors complained to Amazon. Kimberli S. Cox, a breast surgical oncologist based in Arizona, tells The Scientist that she was one of several practicing physicians who in 2008 discovered five-star reviews next to names that were identical or very similar to their own, that they had not written. She and her colleagues successfully persuaded Amazon to take the reviews down. Desai denies that he knew about or was in any way involved in the posting of fake reviews for Surgisphere’s products. “If I wanted to review my own products, I could do it in my own name,” he says. Amazon did not return requests for comment before this story was published. When Desai moved to the University of Texas Health Science Center at Houston in 2012 as a fellow in vascular surgery, he registered Surgisphere Corporation in Texas. By that point, Surgisphere had started publishing the Journal of Surgical Radiology, a medical journal that, according to its website, “accrued over 50,000 subscribers spanning almost every country around the world” from 2010 to 2013. The website further notes that, “With almost one million page views per month, J Surg Rad earned a reputation as one of the first high quality peer-reviewed online medical journals. The Journal was indexed by most of the major medical indexes, and specific articles still appear in PubMed, EBSCO, and other sources.” “It was amazing how fast we were able to grow it,” Desai tells The Scientist. “We had quite the editorial board.” The last issue was published in January 2013. “Running a medical journal is a full-time job,” he says. “I ran out of time.”To recap:
- - Desai says that Surgisphere was always involved with data analytics when it actually started by marketing textbooks.
- -Those products somehow got with fake/fraudulent 5-star reviews on Amazon that impersonated legitimate and prominent doctors.
- - Desai also claims to have started a wildly successful (allegedly) medical journal that was apparently so successful he decided to shut it down.
Surgisphere is currently headquartered in Palatine, Illinois, and run by Desai, who trained in vascular surgery, a subject on which he has published many scientific articles and books. Until February 10 of this year, Desai was employed by Northwest Community Hospital (NCH) in suburban Arlington Heights. He tells The Scientist that he resigned for family reasons. Court records in Cook County, Illinois, show that Desai is named in three medical malpractice lawsuits filed in the second half of 2019. He tells The Scientist in a statement sent through his public relations representative Michael Roth of Bliss Integrated that while he can’t comment on ongoing litigation, he “deems any lawsuit naming him to be unfounded.” He also sent a comment purporting to be from Alan Loren, the executive vice president and chief medical officer of NCH: “Dr. Desai was employed at NCH and resigned in February 2020. We did not have any problems with him while he was here.” Asked by The Scientist if he made this statement, Loren says, “What I can tell you is that he was employed here and he did resign. I can’t speak to whether or not there were any problems.” He adds that he spoke to Desai on May 28 and told him that “what I recall is that he resigned. I don’t remember the exact date. And that was it.”To recap:
- - Has three malpractice suits filed in just the last half of 2019
- - Leaves the associated facility "for family reasons."
- - Puts out a edited/falsified statement from the EVP of the facility that reads "We did not have any problems with him while he was here.” when the actual statement (in pure CYA HR mumbly speak) from the EVP is "I can’t speak to whether or not there were any problems.”