I need your opinion on a new way to design clinical trials for PSP.  New except that it was invented back in the 1980s.  It’s called “combined assessment of function and survival” or CAFS.

When a drug is tested for its neuroprotective property, traditional trial designs have measured only the slowing of the rate of worsening of function (i.e., symptoms and disabilities).  Most of the neuroprotective studies of PSP are 12 months long.  The progression rate of PSP being what it is, about 15% of people with PSP will die each year.  Trial participants who die during a trial are  dropped from the analysis and the data they generated are wasted. 

For PSP and most other chronic, progressive diseases, one hopes that the experimental drug wouldn’t just slow decline of function, but that it would prolong survival as well.  So why not somehow integrate measures of both outcomes into one combined outcome measure?  That could reduce the size of the trials needed to answer the question, which means less expense for the drug company.  That’s important to you and me because it lower the bar for a small, start-up company with a promising drug to mount a clinical trial.  Could the CAFS method work for PSP? 

Combining survival and slowing of progression into one statistical measure would work as follows:  Patients are evenly assigned to receive active drug or placebo.  At the end of the observation period, each patient is compared with each of the other patients, generating either -1 or +1 points for each patient for each comparison.  A patient receives a point if they have survived the observation period but the comparator patient has not.  A patient receives a negative point for the reverse.  If both have died, the patient receives a point if they survived longer than the comparator.  If both have survived (probably the plurality), whoever has progressed more slowly on the standard disability measure over the observation period receives a point.  So, in a study with 50 patients on active drug and 50 on placebo, the best possible score for each patient would be 99, and the worst, -99.  Then, the average points for the patients on active drug is compared with the average for those on placebo.

Sounds great so far, but here’s the rub:  For PSP, the statistics on the progression and survival dictate that this study design assign 50% of the patients to placebo, which is more than the 33% or 25% placebo proportion for most traditional PSP trials.  Furthermore, the observation period would have to be 2 years long, as opposed to the one year of most PSP trials . By that time, your PSP is likely to be too far advanced to qualify for another trial.

That raises some difficult questions for a prospective study subject: What if the drug works but I’m on placebo for the 2 years?  Would converting to the active drug at that later stage of PSP help me?  What if instead I have the opportunity to enroll in a traditionally designed trial  with only a 33% chance of receiving placebo and an observation period of only one year, but with some unattractive features such as greater risk or discomfort or distant and frequent study visits?

So, here are my questions for you. 

Question 1: Suppose that you’re offered a spot in a trial of an experimental drug – let’s call it “PSP1.”  It’s the only trial available to you for the next 6 months, but after that, there may be other trials.  The PSP1 trial uses the novel CAFS design I’ve just described, with a 50% chance of being assigned to placebo and a 2-year duration.  Assume that the science behind the drug suggests that it might slow the rate of PSP progression by 40%, which is quite a large benefit.  Assume that the mode of administration of the drug, the frequency and distance of the study visits and the likelihood of adverse effects are all quite acceptable for you. Would you want to participate despite the possibility that you might receive placebo and that by the end of the 2-year trial you’d be unlikely to benefit much from any neuroprotective drug?

Question 2: Now suppose that 2 drug trials are available to you and, of course, you can join only one.  The first is the PSP1 trial described just above.  The other, for drug “PSP2,” has only the standard 33% chance of placebo and only the standard 12-month duration.  That shorter duration means that you’re likely to still qualify for another drug trial afterwards. Assume the same 40% potential disease-slowing benefit as PSP1.  But the PSP2 trial has some problems such as a greater likelihood of adverse effects, a difficult administration method, or a greater distance or frequency for the study visits. How bad are those problems?  Assume that if the placebo likelihood and the duration were the same as for PSP1, the problems would make PSP2 half as attractive for you as PSP1.

Here’s a convenient summary of all that for you:

	Trial
	PSP1	PSP2
Design	Combined assessment of function and survival (CAFS)	Assessment of function only
Duration	2 years	1 year
Placebo likelihood	50%	33%
Chance of substantial benefit	40%	40%
Practical issues (mode of administration, chance of adverse effects, distance to travel, visit frequency)	Very acceptable	More difficult, making PSP2 half as attractive as PSP1 if all other things were equal.
Likelihood that you could participate in another trial after this	Low	High

OK – time to vote.  Using the Comments feature or ligolbe@gmail.com, please answer these 2 questions:

Question 1:  Would you enroll in the PSP1 trial?  YES or NO

Question 2:  Would you choose the PSP1 trial or the PSP2 trial?

Feel free to include any comments or explanation.