re: survival stats
It is a normal and expected result in studies of terminal diseases that there will be patients who die, and the biostatisticians don’t shy away from that. Unless there are death events to analyse, those stats methods just don’t work.
In very simple terms, the effect of a death event will be balanced out by the extension of life provided by the group of surviving patients.
As pointed out, individual patient updates cannot be statistically or clinically meaningful.
Focusing only on one side of the survival and hazard statistics is meaningless without the comparison to the other side of the equation.
There’s no point discussing the effect of taking one sip from the fluid in the glass, without considering how well the remaining fluid in the glass will last, given such a strong preservative effect with MPL added in.
Once the bottle is opened, obviously the fluid will go off when its time is up, and eventually we will all take our last sip at some point including the patients.
Whilst some amongst the forum crowd might go into instant panic mode upon the news of a death event, scream that part of the sky just fell in, and escalate the drum beat to the point of calling for a coronial inquiry into the cause of patient death, that is dramatic overkill.
The stats are usually called on after a specific amount to time has elapsed, when the proportion of death events are known. Waiting for a certain proportion of death events has the real world problem (in the case of the known efficacy of monepantel) of not knowing when 20%, 30% OR 50% of death events are going to occur. The patient’s extension of life for the group may push to 2, 3 or 5 years. These are unknowable at this point in time.
After the defined time period, the biostatisticians would very calmly and very respectfully, re-calculate both sides of the equation of the analysis. And then, at the appropriate time for a SH update, that aspect of the news will be released.
To put it bluntly, a death event does not mean bad news.
As a statistic, the death event relates to the outcomes of the stats analyses and whether they are statistically significant, and that cannot be evaluated by anyone here without doing the full biostatistical analysis across the entire patient group.
Until that is done, a death event cannot be a material event in and of itself, no matter what personal significance it may have for any individual here. We have to wait for the numbers to come in.
The number of patients may or may not matter or may partially matter. Larger sample sizes usually result in more reliable and accurate estimations. With smaller sample sizes, the estimates may be less precise and more sensitive to individual data points or outliers. Again, we have to wait for the numbers to come in.
Given the remarkable survival and hazard stats provide by monepantel as analysed by Berry Consultants, I’m not expecting any great surprises. The numbers will probably fluctuate as all datasets do over time, but given the increased dosage, all other things being equal, the numbers are more likely to go up, as in beneficial for the patients, rather than down.
cheers,
Ice
(20min delay)