Comparison of in-development treatments for Age Related Macular Degeneration (AMD)
This is a project I did for a statistics course. Thought I would share it on the board.
Objectives:
To conduct an independent unbiased comparison of the effectiveness of Squalamine from Genaera Corporation, an in-development drug for Age Related Macular Degeneration (AMD), to the effectiveness of two competing in-development drugs, Eyetech Pharmaceutical’s Macugen, and Lucentis from Genentech, Inc.
Test to see if data from clinical trials is statistically significant enough to show that Squalamine is more effective than Macugen or Lucentis. If not, test to see if it can be shown that the effectiveness of Squalamine is at least no worse than that of Macugen or Lucentis by some percentage.
Based on differences in clinical data sets, determine how the effectiveness of Squalamine versus Macugen and Lucentis might change in the future.
Introduction:
The data used was taken from publicly available clinical trial results for Macugen, Lucentis, and Squalamine. For Macugen. The clinical data compared was based on availability. Currently, Macugen has completed a Phase III trial, Lucentis and Squalamine have completed Phase II trials. Phase II trails contain a smaller sample and run for less time, while Phase III trials are larger and longer. Squalamine Phase II clinical data was compared to Lucentis Phase I + Phase II data and to Macugen at Phase II and at Phase III. For Squalamine, Phase II data contained a sample of 39 patients for 4 months, Lucentis data was from combined Phase I and Phase II with 53 patients for 2 months. Macugen had 24 patients for 3 months for Phase II and 900 patients for 12 months for Phase III. The significane of the differences in the clinical data is discussed later in this report.
Hypothesis Testing
The available data for each drug was the sample size, the number of patients with improved vision, the number of patients with stable vision, and the number of patients whose vision decreased after the length of the trial. The effectiveness of each drug is better represented by the number of patients whose vision increased. Therefore for this report, patients whose vision increased where considered a success, while patients whose vision was stable or worse were considered a failure. Actual visual increase for each patient was not available. Thus without all of the individual data points the standard deviation of each sample was not able to be calculated. This ruled out any hypothesis test requiring standard deviation. To test for the effectiveness of each drug, the Z test for the differences in two proportions was used. The proportion being compared for each drug was the number of successes (increased vision) over the sample size of the trial.
Tests That Were Run
The Z test for the difference in two proportions takes as parameters:
-The hypothesized difference, a percentage difference between proportion 1 and proportion 2
-The level of significance
-Number of successes and sample size for each population
Six Z tests were run. Two for each comparison:
-Squalamine (13/39) vs. Lucentis (14/53)
-Squalamine (13/39) vs. Macugen (Phase II – 6/24)
-Squalamine (13/39 vs. Macugen (Phase III – 54/900)
95 percent confidence (Level of significance (alpha) = 0.05) was used for all tests.
One-tailed upper tail tests were used for each comparison.
Test 1:
Squalamine (13/39) vs. Macugen (Phase II – 54/900), Hypothesized Difference = 0
This test compares Squalamine to Macugen Phase II data (24 patients, 6 with improved vision). The hypothesized difference is 0. A one-tailed upper-tail Z test is used to determine if Squalamine is more effective than Macugen. The null hypothesis is that the effectiveness of Squalamine is less than or equal to the effectiveness of Macugen. The alternate hypothesis is that Squalamine is more effective than Macugen. Despite Squalamine having a higher observed percentage of patients with improved vision, the data is not statistically significant enough to show that Squalamine is more effective than Macugen. This is because of the very small sample size for both trials. The null hypothesis is not rejected. Because Squalamine can not be shown to be more effective than Macugen, we will test to see if it is no more than 12.5 percent less effective than Macugen.
Test 2:
Squalamine (13/39) vs. Macugen (Phase II – 54/900), Hypothesized Difference = -0.125
This test also compares Squalamine to Macugen Phase II data (24 patients, 6 with improved vision). The hypothesized difference is now –0.125, meaning that the null hypothesis is that the effectiveness of Squalamine is equal to or more than 12.5 percent worse than the effectiveness of Macugen. The alternate hypothesis is that Squalamine is less than 12.5 percent less effective than Macugen. The null hypothesis is rejected. The result of this test shows that we can say with 96 percent certainty (p-value = 0.040) that the effectiveness of Squalamine is less than 12.5 percent worse than that of Macugen.
Test 3:
Squalamine (13/39) vs. Macugen (Phase III – 54/900), Hypothesized Difference = 0
The third test compares Squalamine to the much larger Macugen Phase III clinical data. With the much larger sample size of the Macugen Phase III data it will be easier to show statistically that Squalamine is more effective. The hypothesized difference is 0. The null hypothesis is that Squalamine is equal or less effective than Macugen. The alternate hypothesis is that Squalamine is more effective than Macugen. For this test the null hypothesis is rejected. We can now say with greater than 99 percent certainty that Squalamine is more effective than Macugen. Thus we will test to see if Squalamine is greater than 12.5 percent more effective than Macugen.
Test 4:
Squalamine (13/39) vs. Macugen (Phase III – 54/900), Hypothesized Difference = +0.125
The fourth test is the second test of Squalamine vs. Phase III Macugen data. Because we were able to show that Squalamine was more effective than Macugen with 99 percent certainty, we will now test to see if Squalamine is 12.5 percent more effective than Macugen. The null hypothesis is that Squalamine is less than or equal to 12.5 percent more effective than Macugen. The alternate hypothesis is that Squalamine is greater than 12.5 percent more effective than Macugen. The calculated p-value of 0.0416 tells us that with slightly less than 96 percent certainty we can say that Squalamine is greater than 12.5 percent more effective than Macugen.
Test 5:
Squalamine (13/39) vs. Lucentis (14/53), Hypothesized Difference = 0
The fifth test is between Squalamine and Lucentis. As noted above, Squalamine data is for Phase II, and Lucentis for a combined Phase I and Phase II. The hypothesized difference is 0. The null hypothesis is that the effectiveness of Squalamine is less than or equal to the effectiveness of Lucentis. The alternate hypothesis is that Squalamine is more effective than Lucentis. The p-value of 0.2357 causes us to not reject the null hypothesis. Because Squalamine can not be shown to be statistically more effective than Lucentis, we test to see if Squalamine is no more than 10 percent less effective than Lucentis in the final test, Test 6.
Test 6:
Squalamine (13/39) vs. Lucentis (14/53), Hypothesized Difference = -0.1
As noted, in the final test we will test to see if Squalamine is greater than 10 percent less effective than Lucentis. The hypothesized difference is –0.1. The null hypothesis is that the effectiveness of Squalamine is greater than or equal to 10 percent worse than the effectiveness of Lucentis. The alternate hypothesis is that Squalamine is less than 10 percent worse. The p-value of 0.0391 tells us to reject the null hypothesis and also tells us that with 96 percent confidence we can say that Squalamine is not 10 percent worse than Lucentis.
Test Conclusions:
The result of the hypothesis tests show that when Squalamine is compared to Macugen using the Macugen Phase III data, Squalamine can be shown with 96 percent confidence to be at least 12.5 percent more effective than Macugen. When comparing Squalamine to Macugen’s Phase II data, Squalamine’s better percentage of patients with improved vision (13/39 compared to 6/24) is not statistically significant. Thus it can not be shown that Squalamine is statistically more effective than Macugen when using Phase II data. However it can be shown that Squalamine is no more than 12.5 percent less effective Macugen (96 percent confidence). For Lucentis, no Phase III data is available. Squalamine’s higher percentage of patients with improved vision (13/39 compared to 14/53) is also not statistically significant. It can be shown with 96 percent confidence however that Squalamine is not more than 10 percent less effective than Lucentis.
Report Conclusions:
The tests show that the clinical data for Squalamine is strong. Squalamine’s at Phase II is significantly more effective than Macugen at Phase III. However, as can be seen from Macugen’s data, the effectiveness of a drug can be artificially high in a small Phase II trial. The effectiveness of Squalamine may decrease significantly in a Phase III trial. However no data points to this happening for sure for Squalamine. At this time Squalamine definitely seems to be a stronger drug for the treatment of Age Related Macular Degeneration than Macugen. More data is needed to compare the effectiveness of Squalamine to Lucentis.
Sources:
Genaera Corportation, http://www.genaera.com
Eyetech Pharmaceuticals, http://www.eyetk.com
Genentech, Inc., http://www.gene.com
News 
Market Data 
Discover 
AI
