Rainbow Rare Earths Phalaborwa project shaping up to be one of the lowest cost producers globally. Watch the video here.
HSD, I suppose weâll find out about the vaccination status effect in due course⌠I think steroids would have been SOC at the start of the trial though, havenât they been since summer 2020?
My point was more that (if you look at per protocol group especially) we seem to have gone from âon track for significanceâ at halfway to âalmost no effectâ by the end. Very strange.
Hopefully weâll get some answers soon.
Totally agree Matterhorn - how could we have been on track for significant results on the primary endpoints at halfway, given the results weâve seen from the full cohort? Something doesnât seem to add up.
Cheers guys and glad it helped some. Love a good chi squared test Richlist!
Just on your points Ndn, yes, a larger trial could have swung it the other way, and I shared some of your concerns about the design, but like everyone here I put my faith in the SNG guys on that front. The monetary benefits of a reduction in length of stay in hospital may be easier to reliably quantify, maybe that was a factor in that call, and we donât know what pressure was applied by the government agencies in that regard.
I disagree that itâs rational to look at the P3 results in isolation though, as you seem to be doing, and writing them off. The P2 and P3 results were independent events, and to have gotten p<5% in one and p<12% in another would be very unlikely if the drug has no effect. Because of the SOC difference they canât really be combined though IMO, which is why theyâd need another trial. TW has to choose his words carefully but I think this is what heâs getting at.
Morning all.
Iâve seen some of the conversation around statistical significance for the secondary outcome measures and wanted to share how close I think SNG came to hitting the mark, because I think this could affect how third parties view the data and the drug.
From the data provided in the RNS, we can see that âProgression to severe disease or death within 35 daysâ had an odds ratio of 0.63 for the per-protocol population (i.e. those that had the full course of the drug and did not leave the study early).
This means that for two identical people, one on placebo and one on the drug, the odds of progression to severe disease or death were 37% less for the person on the drug. The probability of getting that result purely by chance was given as 11.9% (we needed to hit 5% or less to be âsignificantâ).
From the data and p-value provided, you can approximate the odds ratio that would have been required to achieve a p-value of 5%, which I have as 0.56, i.e. a 44% reduction in odds (remembering we got 37%).
Doing a little more arithmetic, what this means is that our âback of the netâ p<=5% threshold for the number of per-protocol SNG001 patients progressing to severe disease or death could have been as high as⌠18. We got 20.
If the trial had only been marginally bigger with that 0.63 odds ratio, or if SNG had had slightly better luck with the treatment group, I think weâd have hit that 5% and by my estimation above it was very, very close for that outcome measure on SPRINTER. This was the stand-out outcome measure from the P2 hospital trial as well on an *independent* dataset, a highly noteworthy point IMO.
So in summary, when SNG talk about platform trials, IMO theyâre really not clutching at straws or being desperate. For me, TWâs statement at the end of the RNS is absolutely bang-on.
I couldnât comment on whether or when this will actually happen but it would seem to be in societyâs interest.
For the sake of balance, at face value the primary endpoint data is obviously pants.
Lastly some great perspective from my dad last night: âeverything you own is only really borrowedâ. Not trying to make light of anyoneâs circumstances but it helped me.
Hope the above is useful but as always DYOR and best of luck with whatever you decide to do.
Hi all,
Like many, licking my wounds this morning. Appreciate many have sold but I'm hanging on, rather grimly, for now at least! Hope everyone that's taken a hit is doing ok and is managing to keep a sense of perspective and calm. It's tough to swallow but the best thing you can do IMO is have a chat with someone and share it - picking up the phone certainly helped me this morning.
I just wanted to pick out a line from the RNS, that may indicate some positive news in the short to medium term.
"In the meantime, we eagerly await the Phase 2 data from the US NIH ACTIV-2 trial in home- based COVID-19 patients..."
I know that faith in 'soft signals' from the company will (understandably) be extremely low at present, but since we know this is a trial we have passed, it would seem to suggest positive data could shared with the market in the weeks ahead. Appreciate that this is a line that RM wants us to read an focus on, but all the same would appreciate thoughts from LTH (I know who you are!!)
All the best to LTH today regardless.
Nice one Jake - personally I hadnât seen that Boston branch before, incorporated on Dec 16th - thatâs either day 35 or day 36 after P3 completed recruitment. Odd coincidenceâŚ
Just on Doc83âs point, has everyone read Ray Jordanâs CV? The guy is a world-class professional that has been building his CV and industry reputation for over 40 years. Can you imagine the prospective customers from around the world he *turned down* when launching his business?
Does anyone think itâs realistic that he turned those offers down, and opted for a virtually anonymous ÂŁ350m mcap biotech from Southampton, without getting assurances that P3 was a major success? Do we really think he would have done that just based on RMâs confidence and optimism?
Itâd be like Kevin De Bruyne signing for Hull because the gaffer told him âtrust me son, weâre going places!â
Apologies to any Hull fans on here, but you get my point!
Hi all, thought this might be of interest â the latest (Feb 22) report from NERVTAG on the risk of future Covid variants.
https://www.gov.uk/government/publications/nervtag-long-term-evolution-of-sars-cov-2-10-february-2022
The TL;DR version is (still quite long!):
- âThe next variant to achieve UK/global dominance is likely to have the same pathogenicity as previous variants. The loss of virulence as viruses evolve is a common misconception.â
- Random mutation could increase virulence of existing strains
- NERVTAG are concerned about existing Covid strains recombining in the wild (e.g. Beta and Delta), to create a new Covid strain that is highly infectious, virulent and evades current vaccines (worst of all worlds)
- They are also worried about Covid recombining with an endemic coronavirus, creating a new virus that gets its spike partly from an endemic (unvaccinated against) coronavirus but has the body and virulence of Alpha/Delta.
- Emergence of a strain that leads to âvaccine failureâ is âalmost certainâ but how soon that will happen is unknown.
- They specifically recommend, more than once, to âstockpile prophylactic and therapeutic drugs for SARS-CoV-2â
- It is unlikely in the short term that Covid will follow an evolutionary path of decreasing virulence
- They are concerned about Covid strains that develop a greater ability to suppress the interferon system, which is âa major determinant in disease outcome, individuals deficient in interferon fare badlyâ (para 43)
- It is possible that Covid could develop a degree of interferon-resistance, as this has been seen in other human coronaviruses and revolves around a specific interferon-stimulated gene called OAS1 (para 44)
- A number of concerns are raised around the evolution of strains that are resistant to specific antivirals (like Paxlovid and Molnupiravir)
A bit grim in places but thought it might be a useful reminder for what governments â and by that I mean potential SNG customers â are being advised about the future of the pandemic. Probably a more reliable source than some of the one-post-wonders we get on here! Previous versions of the report may have been shared before but hopefully still helpful/interesting to some.
Morning⌠just to say I completely agree with Bogo there, I could have been clearer, theyâll have had the first draft primary endpoints almost immediately but everything else would and should take more time. Thereâs a lot of analysis left to do after that!
FWIW I subscribe to the long covid theory, as if you release stellar primary endpoint results, that could bias the remaining trial subjects in how they answer the long covid questions and undermine the validity of at least some of that data. If we can raise that concern on this board, Iâm pretty sure a medical regulator would.
Lastly I donât see what SNG have to gain by going public quick or early? Rather than generating press, Iâd rather RM was talking to priority government customers and finalising orders, which is what might be going on right now for all we know.
If I were the US, given the support theyâve lent us Iâd expect first bite at that cherry, before results became public.
Theyâll have looked at all of the day 35 data together in one go after the data were unblinded. Thatâs a requirement for the techniques theyâre using to analyse it.
So after the data were unblinded, they very quickly would have had the results for the full 610.
I think the key point is, once they *knew* whether or not theyâd passed P3, only then did they sign the deal with Ashfield and start recruiting. Perverse thing to do if you know that you have failed.
If they had simply done nothing, itâs not like weâd all be sat here worrying âwhereâs the commercialisation RNS?!â while we wait for P3 readout.
On the balance of probabilities, itâs far, far more likely IMO that P3 has been successful - but not long to wait now regardless.
Just on whether it works - we obviously donât know that for sure yet, but we *do* know that day 35 data were fully unblinded, probably in late Dec. And that since then, SNG have been busy putting in place their commercialisation infrastructure. Rather strange thing to do if it *doesnât* work!
This is section 8.10 of the SG018 trial protocol (kindly shared by another poster):
A primary analysis will take place once all randomised patients have completed their Day 35 visit or have withdrawn from the study. The primary analysis will be used to assess the study data against the study objectives. All data summaries documented in the SAP (Statistical Analysis Plan) will be produced for the primary analysis and will be reported in a CSR (Clinical Study Report), with the exception of any Day 90 summaries. The study will be fully unblinded from the point of the primary analysis. A follow- up analysis will be conducted once all patients have completed the Day 90 visit or have withdrawn from the study.
ThinShins, as far as I can see, the Omicron study author didn't respond to the findings in this pre-print that Delta replication is over 100 times more sensitive to IFN-b relative to IFN-a (fig. 2E):
https://www.biorxiv.org/content/10.1101/2021.11.16.468777v1.full
Which must question how much the effect of IFN-a on Omicron teaches us about the effect of IFN-b (i.e. SNG001) on Omicron.
The new paper also found that there was no difference in IFN-a sensitivity between 'wild type' and Delta, which broadly aligns with what you see in table 2E in the link above. In that paper there was a massive difference in potency between 'wild type' and Delta for IFN-b though, which again must question how much we can learn about IFN-b potency on Omicron from IFN-a data.
Personally, I'll wait to see SNG's own in vitro results before forming any conclusions.
[2 of 2]
3) The hazard ratio is the modelled ratio of the hazards between treatment and placebo. Referring back to the first example, the âhazardâ in question is âare you going to recover todayâ. You may or may not. What a hazard ratio of 1.7 implies is that if you were on SNG001, you were 70% more likely to recover on any given day during the study than if you were in the placebo arm.
The net effect of this is that people on treatment recover sooner, and their average time to recovery is shorter. What a hazard ratio of 1.7 means in those terms is anyoneâs guess â like Matml says, unless youâre Wilko / SSH etc. youâre unlikely to know. Depends on the underlying hazard in the placebo arm. Often *median* differences in time to recovery, time to discharge etc. are quoted to make things more intuitive/relatable, and weâll probably see that in the results RNS (perhaps tomorrow morning?!).
Worth adding that the analyst(s) will be building statistical models of each hazard, which also include the effects of things like age, sex, comorbidities etc. The hazard ratio puts all those to one side, controls for them if you will, and isolates the effect of the treatment vs. the placebo.
For anyone interested, this is called âsurvival analysisâ and this is an awesome link / series of papers if youâre so inclined:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2394262/
4) I think Mary has deducted a 1 from all of his figures so we should read them as hazard ratios of 1.8, 1.5, 1.2 etc. Happy to be corrected here.
Anyways hope the above is helpful and makes things more rather than less clear. If thereâs someone on the board with a medical stats background they might be able to share more, but alas, this is far as my knowledge goes! Have a small child so canât say Iâll be able to reply again tonight but hope this has helped & pays back some of the great research thatâs been shared with me on this board!
All the best for tomorrow everyone!
[1 of 2]
Hi Matterhorn/all, first up thanks for your contributions to the board over the last however many months, I personally really appreciate what youâve shared. Saw this thread last night but it was far too late for me to think about replying then!
I have something of a stats & data science background but medical stats arenât really my thing â having said that, I might be able to offer some assistance. Iâll go through things in order & try to tie things together.
1) Power. With regards to the power of a statistical test, this is an important factor in the design of an experiment / trial. What the âpowerâ gets at is the question âif the true effect size (of the treatment) is as big as I hypothesise, given the number of test subjects, what is the probability that (due to bad luck) I wonât detect any significant effect at all?â In other words, what is the âfalse negativeâ rate. In the case of time to recovery, SNG hypothesise that the hazard ratio (more on that in a sec) for SNG001 with respect to time to recovery is at least 1.7. *If that last part is true*, then with a trial of 610 patients there is a <=10% chance of the p-value of the hazard ratio for that endpoint being >5% in SPRINTER (i.e. the trial not meeting that particular endpoint) just due to bad luck. Equivalently, the trial has 90% power for that endpoint. If the true hazard ratio is actually higher than 1.7 (i.e. SNG001 is more effective than that) then the false negative rate from SPRINTER would in practice have been lower (the trial would have actually had greater than 90% power), even though we couldnât have known that in advance.
So it doesnât mean the meeting the endpoint is 90% likely, it means that *if the effect size is as big as SNG hypothesise* then it is 90% likely. With regards to P3 respiratory trials in general, I think a figure of 70% has recently been stated by FinnCap, but Iâd have it far higher given the futility analysis, Ashfield, and the new jobs & website today!
2) Think Josh has answered this, but in short, a p-value means âif the treatment has no effect, what is the probability that I observed a significant result for that endpoint due to random good luck?â When that is small (typically <5%) we reject the notion that the treatment has no effect on that endpoint. (Ps Josh that coin toss probability starts with a 0. and then has ten zeros before another number!)
Worth adding here that the line âadjusting for the Hochberg procedureâ means that SNG are not going to be dredging the data to try to find a p-value < 5% somewhere. This is important for confidence / validity as even with no effect, if you perform 20 tests, you would expect one of them to return a p-value <= 5% just due to randomness.
It also might suggest that theyâre going to use 110 US patients + the existing UK home trial data to make the decision on progression, as it was supposed to be 220 patients on Activ-2 iirc.
Totally agree Schrow. If they can acquire another 20-30 breathless people over treatment and placebo, and they independently show the same effects as the UK P2, then that will remove any concerns about âdata dredgingâ and quickly change the sentiment I feel. Presumably the A-2 P3 protocol could subsequently be tailored to focus on breathless patients as well...
Hi everyone, given the speculation about the unblinding of the HT data over the last few days, I emailed Consilium to see what I could find. There was a little back and forth required, which I've omitted, but ultimately I received a very helpful reply from Olivia there. In summary, the data will have been unblinded to the statistician (previous poster shared that this is a guy independent from SNG but affiliated with Soton Uni) after day 28 data collection (but when precisely, we don't know), and it was implied that the results would have been shared with RM (as you'd expect) but not the staff 'on the ground' and directly involved in the trial, hence maintaining the double blind for days 60 and 90. Her email below...
-------------------------------------------------------------
Hi xxxx,
Just to follow up on your question below, in a trial like this, where there is a 28 day analysis and a longer endpoint, the data would usually only be unblinded for the 28 day analysis to the stats team (and for example the senior management team) â the study team and site staff still involved in the study would remain double blinded until the whole trial finished. I hope that provides an answer to your question?
Apologies that we are restricted to what is already in the public domain on the finer details of the trial.
Olivia
Just to second a lot of what you're saying Brysoa, to prevent it getting buried. My own research suggested a hospitalisation rate of around 15%, so give or take 9 admissions in placebo, but that is heavily age-dependent and could vary significantly depending on the age of the cohort & the effect of changing hospital pressures over time.
If the treatment group have equivalent ages and comorbidities then off the back of a fag packet, we'll probably need that admissions number to be 4 or less to ping out a p-value below 5%. So it sounds like independently, two people with a background in stats (I'm guessing about you Brysoa!) are essentially suggesting the same thing.
I think as investors we need to be prepared for the outcome that the HT data shows promise without reaching statistical significance, personally that's been my mindset for a while. Like Scinv says, this seems to be a very small trial with seemingly low power to detect an effect once you plug in the hospitalisation rates.
I'm not sure I'm 100% behind the hypothesis that the trial has been changed in response to initial unblinding though. If the data are to be amalgamated with Activ-2, and the US is to be our principal customer, then what need do SNG have to publish data now? And if the data were negative, wouldn't RM have a responsibility to inform the market that the drug had failed to meet the primary endpoint, which remains only up to day 28 in the revised trial protocol? Happy to be corrected (politely) if I'm wrong on that point.
I think so Graham, I read somewhere on the conference site that the videos would be accessible to attendees until the 7th/8th of April. So I think theyâre just putting it up on YouTube and removing the paywall on that date.
Iâll still be watching when I can, could be interesting and grateful to the people that found it for sharing.