Let’s Break it Down-3 (Tipping the Scales)

So, here we go. I struggled with this section because of the graphs and numbers. There may be many of you who go…”What’s the big deal, so easy to understand!” Me, to say I’m good at math would be stretch and I wanted to be sure that I understood. So, I went to a researcher, sweet man gave me SO much more credit of possibly understanding his answer than I deserved. Thank you, you know who you are! Not wanting to bother him again (or show my severe lack of understanding to someone I respect)  next, I tried a statistic’s teacher that I did a trial presentation for, no answer. Undeterred, I went straight to the UCLA Statistics Dept. and after some back and forth with Dr. Li, Assistant Professor Applied Statistics and Statistical Modeling, as well as their interface with Statistical Genomics, Bioinformatics, and Computational Biology, I came to have a very basic understanding, but one she said was correct! So, now on to Part 3!

Key outcomes from the trial:
82 subjects were randomized into the following groups: 24 to placebo BID, 27 to 200mg/kg BID, 16 to 100mg/kg BID and 15 to 50mg/kg BID.  The mean and median age, weight and BMI did not differ significantly between the four groups.
The pre-specified efficacy analyses prioritized 5 core syndrome-specific measures:

1. The RSBQ, a rating scale in which the subject’s caregiver rates the frequency of symptoms.

 2. The CGI-I, in which the clinician rates how much the subject’s overall illness has improved or worsened, relative to baseline. 

3. The RTT-DSC, in which the clinician assesses on a visual analog scale the severity of concerns identified for each subject on an individual basis. 

4. The Motor Behavior Assessment (MBA), a rating scale in which the clinician rates the subject’s current level of function.   

5. The Caregiver Top 3 Concerns (Top 3), in which the subject’s caregiver assesses on a visual analog scale the severity of concerns identified for each subject on an individual basis.

The first three of these measures each demonstrated improvement for the highest dose, compared with placebo, which was both statistically significant and clinically meaningful.  The improvement increased throughout the course of treatment right through to the time that treatment ceased.  This suggests that further benefit may be achieved with even longer treatment duration.  Some regression of benefit was observed after treatment ceased.  These results are illustrated in the following charts, in which a downward movement represents an improvement from day 14 baseline:  (graphs should be here, but I think they would just confuse the issue and I couldn’t get them to transfer over, but you can view them on page 3-4 of Neuren’s announcement)

22% of subjects in the 200mg/kg dose group received a CGI-I score of 2 (“much improved”) compared with 4% of subjects in the placebo group.

The mean (lsmean) improvements at day 54 for the 200mg/kg and placebo groups were, respectively, 16% and 6% of the day 14 baseline.

(Only the first three core measures will be discussed in this post) So, here we go…
82 subjects were randomized into the following groups: 24 to placebo BID, 27 to 200mg/kg BID, 16 to 100mg/kg BID and 15 to 50mg/kg BID.  The mean and median age, weight and BMI did not differ significantly between the four groups.

Easy! Of 82 participants-24 got the placebo twice a day; 27 got 200mg/kg (200mg per kilogram a kilogram being 2.2 lbs) twice a day; 16 got 100mg/kg twice a day and 15 got 50mg/kg twice a day. The rest means that if you averaged in age, weight, and BMI (Body Mass Index) it didn’t really favor one group over the other.

The pre-specified efficacy analyses prioritized 5 core syndrome-specific measures:

Pre-specified efficacy means that before the trial even started they set goals which, if reached, would show efficacy.

1. The RSBQ, a rating scale in which the subject’s caregiver rates the frequency of symptoms.

The Rett Syndrome Behavior Questionnaire is a recognized standard of 45 questions rated by the parent.

 2. The CGI-I, in which the clinician rates how much the subject’s overall illness has improved or worsened, relative to baseline. 

The Clinical Global Impression – Improvement scale (CGI-I) is a 7 point scale that requires the clinician to assess how much the participant’s illness has improved or worsened relative to a baseline state at the beginning of the intervention. and rated as: 1, very much improved; 2, much improved; 3, minimally improved; 4, no change; 5, minimally worse; 6, much worse; or 7, very much worse.

3. The RTT-DSC (Rett Syndrome Domain Specific Concerns Visual Analog Scale), in which the clinician assesses on a visual analog scale the severity of concerns identified for each subject on an individual basis. Discussed in the last post

The first three of these measures each demonstrated improvement for the highest dose, compared with placebo, which was both statistically significant and clinically meaningful.  The improvement increased throughout the course of treatment right through to the time that treatment ceased.  This suggests that further benefit may be achieved with even longer treatment duration.  Some regression of benefit was observed after treatment ceased.  These results are illustrated in the following charts, in which a downward movement represents an improvement from day 14 baseline: 

So, this paragraph is GREAT NEWS! The RSBQ score, the CGI-I score and the RTT-DSC score all showed improvement. The improvement increased throughout and up to the time treatment ceased; which means that a cap on improvement was not discovered and that longer duration of treatment MAY see an even greater improvement in function. It also shows that, without Trofinetide, regression to baseline began (and presumably eventually the children returned to baseline) While this really sucks, it is still a great indicator that Trofinetide worked.

The mean (lsmean) improvements at day 54 for the 200mg/kg and placebo groups were, respectively, 16% and 6% of the day 14 baseline.

Now, for the part I was stuck on. What is an lsmean?? and how could a numeric decrease translate into a % increase. Turns out lsmean is a fairly complicated calculation of least squares mean and in my opinion and determination the exactness of how this is achieved is not relevant to our understanding of the results.

The graphs (not pictured) show a negative number and yet indicate a substantial improvement. So, this was the question I posed to Dr. Li….

So, IF a numeric scale is used to show a negative effect for example “severity of screaming” and that initially gets a “2” but then goes DOWN to “1” this is a DECREASE in measurement that ultimately shows an INCREASE from baseline toward “normal”?

scales-of-justice-clip-art-lady-tattoo-clipart-best-clipart-best-Xq22G6-clipartAnd she said that my understanding was correct. That is why I titled this post “Tipping the scales”. Picture the Scales of Justice, as one side goes down… the other side goes up, finally in our favor!

And you may ask why compared to day 14, because that was how long the placebo lead in, in all groups, lasted and is the true baseline.

22% of subjects in the 200mg/kg dose group received a CGI-I score of 2 (“much improved”) compared with 4% of subjects in the placebo group. This is self-explanatory.

The group analysis of RTT-DSC was carried out using the Exact Median Test rather than the Least squares Means that were calculated for the other four core endpoints.  This was because the Statistical Analysis Plan pre-specified that if the data for a measure did not meet statistical assumptions of a general linear model, then a non-parametric analysis method would be used.  Standard error limits are not applicable in the Exact Median Test and consequently are not presented on the bar chart.
The median improvements at day 54 for the 200mg/kg and placebo groups were, respectively, 15% and 5% of the day 14 baseline.

Basically, the RTT-DSC used a different measurement because it could not correctly be represented by the other. Regardless, it shows a similar percentage of improvement for both the placebo and the 200mg/kg dose of Trofinetide.

Whoa! Glad that part is over!!! If there is anyone out there who can explain it better, please let me know and I will happily defer and edit!

Advertisements
This entry was posted in Neuren Pharmaceuticals, Trail to a Texas Trial, Trofinetide, Trofinetide Pediatric trial, ucla statistics dept., Uncategorized and tagged , , , , , . Bookmark the permalink.

2 Responses to Let’s Break it Down-3 (Tipping the Scales)

  1. Rose and Renée says:

    I hope everyone with interest in this clinical trial sees your layman’s terms. This is an exceptionally huge help.
    Thanks Mel!

    • melelllan says:

      Thank you, Rosemary. It’s been an exercise in true patience with myself searching for the right words and examples/analogies that are applicable. Sometimes I read it and wonder if it makes sense. I’ll be happy when I get through the whole thing. These graphs and math caused me some sleepless nights! Lol

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s