1
00:00:07,569 --> 00:00:08,125
>> GINA: WELCOME TO
EVERYONE 

2
00:00:08,125 --> 00:00:11,095
WHO HAS JOINED TODAY, EITHER ON 

3
00:00:11,095 --> 00:00:13,597
ZOOM OR ON VIDEOCAST. WELCOME 

4
00:00:13,597 --> 00:00:14,164
TO OUR FIRST OF SEVERAL 

5
00:00:14,164 --> 00:00:24,341
LECTURES FOR THE ANNUAL NINDS 

6
00:00:24,341 --> 00:00:24,942
SERIES ON STATISTICAL SERIES 

7
00:00:24,942 --> 00:00:29,380
AND I AM GINA NORATO.  I WILL 

8
00:00:29,380 --> 00:00:31,248
BE TALKING THIS WEEK ABOUT 

9
00:00:31,248 --> 00:00:31,815
GIVING A BRIEF OVERVIEW FOR 

10
00:00:31,815 --> 00:00:36,587
CLINICAL TRIALS DESIGN. 

11
00:00:36,587 --> 00:00:38,622
ON WHATEVER PLATFORM YOU ARE 

12
00:00:38,622 --> 00:00:39,456
ON, PLEASE FEEL FREE TO SEND US 

13
00:00:39,456 --> 00:00:42,826
QUESTIONS. AND I CAN TRY TO 

14
00:00:42,826 --> 00:00:47,164
ADDRESS THEM AS WE GO.  IF YOU 

15
00:00:47,164 --> 00:00:47,664
FEEL COMFORTABLE ASKING 

16
00:00:47,664 --> 00:00:48,265
QUESTIONS VERBALLY, FEEL FREE 

17
00:00:48,265 --> 00:00:49,433
TO DO THAT.  THIS WILL BE 

18
00:00:49,433 --> 00:00:56,640
RECORDED. SO KEEP THAT IN MIND. 

19
00:00:56,640 --> 00:01:00,511
SO, THE FIRST THING THAT I WANT 

20
00:01:00,511 --> 00:01:02,045
TO DO IS GIVE A GENERAL OUTLINE 

21
00:01:02,045 --> 00:01:02,613
FOR WHAT WE ARE GOING TO BE 

22
00:01:02,613 --> 00:01:07,785
TALKING ABOUT. 

23
00:01:07,785 --> 00:01:08,385
FIRSTLY, I WANT TO GO OVER THIS 

24
00:01:08,385 --> 00:01:10,320
GENERALLY WHAT A CLINICAL TRIAL 

25
00:01:10,320 --> 00:01:14,725
IS. THE DIFFERENT PHASES OF 

26
00:01:14,725 --> 00:01:15,292
CLINICAL TRIALS AND ETHICAL 

27
00:01:15,292 --> 00:01:15,926
CONSIDERATIONS YOU SHOULD THINK 

28
00:01:15,926 --> 00:01:16,560
ABOUT YOU CONSIDER WHETHER OR 

29
00:01:16,560 --> 00:01:17,194
NOT CLINICAL TRIAL IS RIGHT FOR 

30
00:01:17,194 --> 00:01:19,329
YOU.  AND HOW TO START BUILDING 

31
00:01:19,329 --> 00:01:23,066
IT. 

32
00:01:23,066 --> 00:01:23,534
WE GOING TO GO INTO KEY 

33
00:01:23,534 --> 00:01:27,471
PRINCIPLES OF TRIAL DESIGN, AND 

34
00:01:27,471 --> 00:01:27,971
THINGS YOU NEED TO MAKE 

35
00:01:27,971 --> 00:01:28,605
DECISIONS ON AS YOU DESIGN YOUR 

36
00:01:28,605 --> 00:01:30,707
TRIAL.  THOSE ARE GOING TO BE 

37
00:01:30,707 --> 00:01:38,315
CHOICE OF CONTROL, CONTROL ARMS.

38
00:01:38,315 --> 00:01:44,621
 CHOICE OF ENDPOINTS, CONTROL 

39
00:01:44,621 --> 00:01:51,028
OF BIAS AND COMMON DESIGNS.

40
00:01:51,028 --> 00:01:51,595
AND ALSO GOING TO TALK ABOUT 

41
00:01:51,595 --> 00:01:52,196
STATISTICAL ANALYSES AND THE 

42
00:01:52,196 --> 00:01:52,763
ROLE OF THE STATISTICIAN IN 

43
00:01:52,763 --> 00:01:54,398
THIS PROCESS.  THIS IS GOING TO 

44
00:01:54,398 --> 00:01:58,001
BE REALLY A BROAD OVERVIEW.  

45
00:01:58,001 --> 00:01:58,602
EACH OF THESE THINGS COULD BE 

46
00:01:58,602 --> 00:02:01,004
GIVEN A FULL HOUR LECTURE.  

47
00:02:01,004 --> 00:02:01,572
THIS IS GOING TO BE KIND OF 

48
00:02:01,572 --> 00:02:03,540
CLINICAL TRIALS INTRO CRASH 

49
00:02:03,540 --> 00:02:06,210
COURSE.  SO HOPEFULLY THIS 

50
00:02:06,210 --> 00:02:09,112
GIVES EVERYBODY WHO'S LISTENING 

51
00:02:09,112 --> 00:02:12,950
AN IDEA OF WHAT ULTIMATELY IN A 

52
00:02:12,950 --> 00:02:15,419
LOT OF WAYS ARE STATISTICAL 

53
00:02:15,419 --> 00:02:18,222
SORT OF DECISIONS ABOUT BIAS, 

54
00:02:18,222 --> 00:02:20,858
AND DECISION-MAKING, FOR 

55
00:02:20,858 --> 00:02:21,959
ANALYTICAL COMPARISONS AND 

56
00:02:21,959 --> 00:02:23,760
DEFINITION OF ENDPOINTS. AND 

57
00:02:23,760 --> 00:02:24,795
GIVING US A GOOD FRAMEWORK FOR 

58
00:02:24,795 --> 00:02:33,837
US TO START FROM. 

59
00:02:33,837 --> 00:02:34,438
SO, AGAIN FIRST I WANT TO TALK 

60
00:02:34,438 --> 00:02:39,676
ABOUT A BROAD OVERVIEW.  NIH 

61
00:02:39,676 --> 00:02:43,747
GRANTS AND FUNDING DESIGNS  

62
00:02:43,747 --> 00:02:44,348
CLINICAL TRIALS AS A RESEARCH 

63
00:02:44,348 --> 00:02:45,682
STUDY IN WHICH ONE OR MORE 

64
00:02:45,682 --> 00:02:46,116
HUMAN SUBJECTS ARE 

65
00:02:46,116 --> 00:02:47,551
PROSPECTIVELY ASSIGNED TO ONE 

66
00:02:47,551 --> 00:02:49,219
OR MORE INTERVENTIONS, WHICH 

67
00:02:49,219 --> 00:02:52,122
MAY INCLUDE PLACEBO OR OTHER 

68
00:02:52,122 --> 00:02:52,656
CONTROL, TO EVALUATE THE 

69
00:02:52,656 --> 00:02:53,390
EFFECTS OF THOSE INTERVENTIONS 

70
00:02:53,390 --> 00:02:54,725
ON HEALTH RELATED BIOMEDICAL OR 

71
00:02:54,725 --> 00:03:00,030
BEHAVIORAL OUTCOMES. 

72
00:03:00,030 --> 00:03:02,599
AND THEY ALSO PROVIDE KIND OF A 

73
00:03:02,599 --> 00:03:07,938
HELPFUL SET OF QUESTIONS.  SO 

74
00:03:07,938 --> 00:03:08,572
YOU CAN DETERMINE IF YOUR STUDY 

75
00:03:08,572 --> 00:03:10,307
IS A CLINICAL TRIAL.  THIS IS 

76
00:03:10,307 --> 00:03:11,909
GOING TO BE IMPORTANT, NOT ONLY 

77
00:03:11,909 --> 00:03:13,110
FOR ANALYTICAL REASONS, BUT FOR 

78
00:03:13,110 --> 00:03:17,080
REPORTING REASONS. IF YOUR 

79
00:03:17,080 --> 00:03:18,382
STUDY IS A CLINICAL TRIAL WE 

80
00:03:18,382 --> 00:03:19,683
HAVE TO REPORT TO CLICK THE 

81
00:03:19,683 --> 00:03:23,620
TRIALS DOT GOV, MANAGED IN 

82
00:03:23,620 --> 00:03:27,157
OTHER WAYS, MORE RIGOROUS THAN 

83
00:03:27,157 --> 00:03:27,691
OTHER SORTS OF TRIALS OR 

84
00:03:27,691 --> 00:03:28,592
DESIGNS. 

85
00:03:28,592 --> 00:03:33,263
YOU CAN ASK YOURSELF,  DOES THE 

86
00:03:33,263 --> 00:03:33,697
STUDY INVOLVE HUMAN 

87
00:03:33,697 --> 00:03:34,164
PARTICIPANTS? ARE THE 

88
00:03:34,164 --> 00:03:35,198
PARTICIPANTS PROSPECTIVELY 

89
00:03:35,198 --> 00:03:35,866
ASSIGNED TO AN INTERVENTION? IS 

90
00:03:35,866 --> 00:03:37,634
THE STUDY DESIGNED TO EVALUATE 

91
00:03:37,634 --> 00:03:39,803
THE EFFECTS OF THE INTERVENTION 

92
00:03:39,803 --> 00:03:46,376
ON THE PARTICIPANTS? AND, 

93
00:03:46,376 --> 00:03:46,977
FINALLY, IS THE EFFECT BEING 

94
00:03:46,977 --> 00:03:47,511
EVALUATED HEALTH-RELATED 

95
00:03:47,511 --> 00:03:48,045
BIOMEDICAL OR BEHAVIORAL 

96
00:03:48,045 --> 00:03:53,150
OUTCOME?

97
00:03:53,150 --> 00:03:55,819
AND THERE ARE GENERALLY FOUR 

98
00:03:55,819 --> 00:03:57,888
PHASES OF CLINICAL TRIALS.  

99
00:03:57,888 --> 00:04:01,792
PRIOR TO THIS, THE DRUG WOULD 

100
00:04:01,792 --> 00:04:05,262
BE IN WHAT IS CALLED IN THE 

101
00:04:05,262 --> 00:04:09,066
ZERO OR PRECLINICAL FACES, 

102
00:04:09,066 --> 00:04:09,700
WHERE IT WOULD BE APPROVED FOR 

103
00:04:09,700 --> 00:04:14,871
FURTHER HUMAN TESTING. PHASE 1 

104
00:04:14,871 --> 00:04:18,408
ASKS ABOUT THE SAFETY. SIDE 

105
00:04:18,408 --> 00:04:19,042
EFFECTS, HOW DOES IT AFFECT OR 

106
00:04:19,042 --> 00:04:29,419
MOVE THE BODY? IS A SAFETY USE 

107
00:04:29,419 --> 00:04:31,421
AT THE SAME TIME AS OTHER 

108
00:04:31,421 --> 00:04:32,055
MEDICATIONS? GENERALLY THIS IS 

109
00:04:32,055 --> 00:04:32,656
DONE ON A VERY SMALL GROUP OF 

110
00:04:32,656 --> 00:04:33,490
PEOPLE. USUALLY DONE IN HEALTHY 

111
00:04:33,490 --> 00:04:33,957
PEOPLE FOR AFFECTED 

112
00:04:33,957 --> 00:04:40,330
INDIVIDUALS.  ALSO DONE IN THE 

113
00:04:40,330 --> 00:04:40,864
DISEASE GROUP OF INTEREST 

114
00:04:40,864 --> 00:04:42,699
DEPENDING ON THE AVAILABLE 

115
00:04:42,699 --> 00:04:45,769
COHORT AND THE SPECIFIC GOALS.  

116
00:04:45,769 --> 00:04:46,403
AND WHETHER OR NOT THIS KIND OF 

117
00:04:46,403 --> 00:04:48,939
LEADS INTO BEING PHASE 1, PHASE 

118
00:04:48,939 --> 00:04:51,041
2 STUDY. 

119
00:04:51,041 --> 00:04:54,444
PHASE 2 IS MORE INTERESTED IN 

120
00:04:54,444 --> 00:04:54,978
ANSWERING QUESTIONS ABOUT 

121
00:04:54,978 --> 00:04:58,081
EFFICACY.  SO WHILE WE STILL 

122
00:04:58,081 --> 00:05:02,452
ASK QUESTIONS ABOUT SAFETY, AND 

123
00:05:02,452 --> 00:05:03,086
REVIEW THROUGHOUT THE DIFFERENT 

124
00:05:03,086 --> 00:05:05,022
PHASES OF CLINICAL TRIAL 

125
00:05:05,022 --> 00:05:05,622
DESIGN, ADVOCACY IS A LITTLE 

126
00:05:05,622 --> 00:05:06,156
BIT MORE ANSWERED AT THIS 

127
00:05:06,156 --> 00:05:10,961
STAGE.  IT IS MORE COMMONLY 

128
00:05:10,961 --> 00:05:12,896
USING THE ACTUAL POPULATION OF 

129
00:05:12,896 --> 00:05:13,530
INTEREST, OR INVESTIGATING THE 

130
00:05:13,530 --> 00:05:15,499
CONDITION BEING STUDIED. SO 

131
00:05:15,499 --> 00:05:18,335
ASKING QUESTIONS ABOUT, DOES 

132
00:05:18,335 --> 00:05:25,442
THE TREATMENT OF INTEREST  

133
00:05:25,442 --> 00:05:26,076
RELIEVE OR STOP THE PROGRESSION 

134
00:05:26,076 --> 00:05:26,610
OF THE CONDITION? ASKING 

135
00:05:26,610 --> 00:05:27,177
QUESTIONS ABOUT USAGE WOULD 

136
00:05:27,177 --> 00:05:28,445
ALSO BE A GOOD PHASE TO 

137
00:05:28,445 --> 00:05:29,079
QUESTION. GENERALLY IN A LARGER 

138
00:05:29,079 --> 00:05:33,984
SAMPLE.  AND THE NUMBERS ARE 

139
00:05:33,984 --> 00:05:34,551
GOING TO VARY, THE NUMBERS 

140
00:05:34,551 --> 00:05:36,386
LISTED ON THIS FIGURE. THESE 

141
00:05:36,386 --> 00:05:39,322
ARE GOING TO BURY BASED ON YOUR 

142
00:05:39,322 --> 00:05:41,825
AVAILABLE COHORT AND FOR US, WE 

143
00:05:41,825 --> 00:05:45,195
SEE A LITTLE BIT SMALLER THAN 

144
00:05:45,195 --> 00:05:45,829
THIS BUT TO GIVE YOU AN IDEA OF 

145
00:05:45,829 --> 00:05:47,497
HOW CAN IT CAN ESCALATE AS YOU 

146
00:05:47,497 --> 00:05:51,768
PROCEED THROUGH THE PHASES. 

147
00:05:51,768 --> 00:05:53,537
PHASE 3 IS CONSIDERED MORE OF A 

148
00:05:53,537 --> 00:05:59,443
CONFIRMATORY PHASE OF TRIAL.  

149
00:05:59,443 --> 00:06:03,180
FURTHER EFFECTIVENESS, AND 

150
00:06:03,180 --> 00:06:06,683
ADVERSE EFFECTS, ASKING BROADER 

151
00:06:06,683 --> 00:06:07,217
QUESTIONS ABOUT SURVIVAL, 

152
00:06:07,217 --> 00:06:10,287
LONG-TERM QUESTIONS, QUALITY OF 

153
00:06:10,287 --> 00:06:11,722
LIFE, LONG-TERM ADVERSE 

154
00:06:11,722 --> 00:06:15,425
EFFECTS, AND AGAIN LARGER 

155
00:06:15,425 --> 00:06:16,393
SAMPLE SIZE. BEING ABLE TO 

156
00:06:16,393 --> 00:06:17,761
REACH MORE QUESTIONS, WITH 

157
00:06:17,761 --> 00:06:21,865
LARGER SAMPLES. 

158
00:06:21,865 --> 00:06:22,399
AND PHASE 4 IS GOING TO BE 

159
00:06:22,399 --> 00:06:23,467
ABOUT AS BIG AS YOU CAN GET IN 

160
00:06:23,467 --> 00:06:31,441
TERMS OF LOOKING AT THE 

161
00:06:31,441 --> 00:06:32,309
INVESTIGATIONAL PRODUCT. IN A 

162
00:06:32,309 --> 00:06:33,777
VERY LARGE COHORT, OVER LONG 

163
00:06:33,777 --> 00:06:36,313
PERIODS OF TIME.  TO ANSWER 

164
00:06:36,313 --> 00:06:38,582
QUESTIONS ABOUT VERY LONG-TERM 

165
00:06:38,582 --> 00:06:41,218
RISKS AND BENEFITS AND OFTEN 

166
00:06:41,218 --> 00:06:42,686
THIS CAN BE AFTER THE DRUG IS 

167
00:06:42,686 --> 00:06:48,558
APPROVED.  COLLECTING FURTHER 

168
00:06:48,558 --> 00:06:49,159
INFORMATION AS PEOPLE USE THE 

169
00:06:49,159 --> 00:06:49,760
DRUG, IN KIND OF A FOLLOW-UP 

170
00:06:49,760 --> 00:06:55,565
PHASE. 

171
00:06:55,565 --> 00:06:57,834
SO, AS YOU DESIGN, AS YOU START 

172
00:06:57,834 --> 00:07:00,003
TO THINK ABOUT DESIGNING, 

173
00:07:00,003 --> 00:07:02,739
THERE'S GOING TO BE A HANDFUL 

174
00:07:02,739 --> 00:07:03,340
OF KEY ETHICAL CONSIDERATIONS 

175
00:07:03,340 --> 00:07:03,974
YOU NEED TO THINK ABOUT AS YOU 

176
00:07:03,974 --> 00:07:06,810
DESIGN THE TRIAL.  AND YOU 

177
00:07:06,810 --> 00:07:07,377
REALLY WANT TO THINK ABOUT 

178
00:07:07,377 --> 00:07:10,247
THESE EARLY ON AS WELL.  TO 

179
00:07:10,247 --> 00:07:10,847
ENSURE YOU ARE DOING THIS IN 

180
00:07:10,847 --> 00:07:12,983
THE BEST WAY POSSIBLE.

181
00:07:12,983 --> 00:07:13,583
FIRST IS GOING TO BE SOCIAL AND 

182
00:07:13,583 --> 00:07:16,153
CLINICAL VALUE. YOU REALLY WANT 

183
00:07:16,153 --> 00:07:17,854
TO MAKE SURE THAT THE RESEARCH 

184
00:07:17,854 --> 00:07:19,256
QUESTION IS IMPORTANT OR 

185
00:07:19,256 --> 00:07:19,823
VALUABLE ENOUGH TO JUSTIFY 

186
00:07:19,823 --> 00:07:23,493
ASKING PEOPLE TO TAKE RISKS, OR 

187
00:07:23,493 --> 00:07:24,127
INCONVENIENCE TO PARTICIPATE IN 

188
00:07:24,127 --> 00:07:27,130
YOUR STUDY.  THE STUDY SHOULD 

189
00:07:27,130 --> 00:07:27,664
CONTRIBUTE TO SCIENTIFIC 

190
00:07:27,664 --> 00:07:29,166
UNDERSTANDING AND IMPROVE THE 

191
00:07:29,166 --> 00:07:29,733
TREATMENT AND PREVENTION OF 

192
00:07:29,733 --> 00:07:31,768
DISEASE.

193
00:07:31,768 --> 00:07:35,238
NEXT SCIENTIFIC VALIDITY.  WE 

194
00:07:35,238 --> 00:07:35,872
WANT TO MAKE SURE WE CAN ANSWER 

195
00:07:35,872 --> 00:07:36,506
THE RESEARCH QUESTION THAT WE 

196
00:07:36,506 --> 00:07:40,277
ARE TRYING TO ASK.  ARE THE 

197
00:07:40,277 --> 00:07:40,877
METHODS VALID AND FEASIBLE TO 

198
00:07:40,877 --> 00:07:42,179
CONDUCT? IS THERE A VERY CLEAR 

199
00:07:42,179 --> 00:07:45,515
OBJECTIVE, AND VALIDATED 

200
00:07:45,515 --> 00:07:46,116
PRINCIPLES FOR CONDUCTING THE 

201
00:07:46,116 --> 00:07:53,824
STUDY?THUS THE STATISTICAL PLAN 

202
00:07:53,824 --> 00:07:59,196
OUTLINED AND WELL POWERED?  SO 

203
00:07:59,196 --> 00:07:59,796
WE CAN MAKE SURE THE STUDY IS 

204
00:07:59,796 --> 00:08:00,397
DESIGNED IN A SCIENTIFICALLY 

205
00:08:00,397 --> 00:08:04,167
VALID WAY. MINIMIZING RESEARCH 

206
00:08:04,167 --> 00:08:08,638
WAYS AND ALL THESE THINGS. 

207
00:08:08,638 --> 00:08:09,206
AND THEN WE HAVE FAIR SUBJECT 

208
00:08:09,206 --> 00:08:12,542
SELECTION.  WE SHOULD BE 

209
00:08:12,542 --> 00:08:13,109
CAREFUL TO CHOOSE PEOPLE IN 

210
00:08:13,109 --> 00:08:13,877
WITH MINIMIZE RISKS AND 

211
00:08:13,877 --> 00:08:14,811
ENHANCES BENEFITS TO 

212
00:08:14,811 --> 00:08:17,347
INDIVIDUALS AND SOCIETY.  AND 

213
00:08:17,347 --> 00:08:21,051
AS MANY OF US KNOW WHO HAVE 

214
00:08:21,051 --> 00:08:21,618
GONE THROUGH THESE TYPES OF 

215
00:08:21,618 --> 00:08:23,253
TRAININGS BEFORE, SPECIFIC 

216
00:08:23,253 --> 00:08:23,854
GROUPS SHOULD NOT BE EXCLUDED 

217
00:08:23,854 --> 00:08:24,454
FROM THE OPPORTUNITY TO 

218
00:08:24,454 --> 00:08:24,988
PARTICIPATE WITHOUT GOOD 

219
00:08:24,988 --> 00:08:29,392
REASON. 

220
00:08:29,392 --> 00:08:29,926
WE SHOULD HAVE A FAVORABLE 

221
00:08:29,926 --> 00:08:33,964
RISK-BENEFIT RATIO.  TO MAKE 

222
00:08:33,964 --> 00:08:35,432
SURE EVERYTHING HAS BEEN DONE 

223
00:08:35,432 --> 00:08:36,199
TO MINIMIZE RISK, AND MAXIMIZE 

224
00:08:36,199 --> 00:08:36,833
THE POTENTIAL BENEFITS THAT WE 

225
00:08:36,833 --> 00:08:38,969
MIGHT SEE. 

226
00:08:38,969 --> 00:08:39,669
WE SHOULD HAVE SOME INDEPENDENT 

227
00:08:39,669 --> 00:08:44,941
REVIEW. SO ARE THOSE 

228
00:08:44,941 --> 00:08:45,508
INDIVIDUALS, THE RESEARCH, 

229
00:08:45,508 --> 00:08:49,212
STATISTICIANS, STUDY TEAM, ARE 

230
00:08:49,212 --> 00:08:49,846
THEY SUFFICIENTLY FREE OF BIAS? 

231
00:08:49,846 --> 00:08:51,648
OR ARE WE TAKING ALL THE STEPS 

232
00:08:51,648 --> 00:08:53,049
TO MAKE SURE OUR STUDY HAS BEEN 

233
00:08:53,049 --> 00:09:03,093
LOOKED AT BY OTHER SOURCES? AND 

234
00:09:03,093 --> 00:09:03,660
THIS GOES THROUGH A REVIEW 

235
00:09:03,660 --> 00:09:04,394
PROCESS, BUILT IN TWO MOST OF 

236
00:09:04,394 --> 00:09:08,531
THE REPORTS AT NIH. 

237
00:09:08,531 --> 00:09:09,599
WE ALSO NEED TO MAKE SURE THAT 

238
00:09:09,599 --> 00:09:11,568
WE ARE ADHERING TO GOOD, 

239
00:09:11,568 --> 00:09:15,338
INFORMED CONSENT PRINCIPLES.  

240
00:09:15,338 --> 00:09:15,906
SO THE INDIVIDUALS MUST BE 

241
00:09:15,906 --> 00:09:16,439
INFORMED OF THE PURPOSE, 

242
00:09:16,439 --> 00:09:17,641
METHODS, RISKS, BENEFITS AND 

243
00:09:17,641 --> 00:09:19,809
ALTERNATIVES TO THE STUDY.  AND 

244
00:09:19,809 --> 00:09:20,443
THEY REALLY NEED TO BE ABLE TO 

245
00:09:20,443 --> 00:09:23,513
MAKE THEIR OWN EDUCATED 

246
00:09:23,513 --> 00:09:24,080
DECISIONS ABOUT WHETHER TO 

247
00:09:24,080 --> 00:09:29,653
PARTICIPATE.

248
00:09:29,653 --> 00:09:33,823
AND FINALLY, WE NEED TO RESPECT 

249
00:09:33,823 --> 00:09:34,357
POTENTIAL AND AND ROLLED 

250
00:09:34,357 --> 00:09:34,824
SUBJECTS AND RESPECT 

251
00:09:34,824 --> 00:09:39,663
CONFIDENTIALITY.  PATIENTS 

252
00:09:39,663 --> 00:09:40,263
SHOULD BE ALLOWED TO WITHDRAW 

253
00:09:40,263 --> 00:09:43,233
FROM THE STUDY AT ANY POINT. 

254
00:09:43,233 --> 00:09:43,800
PARTICIPANTS ALSO HAVE THE 

255
00:09:43,800 --> 00:09:44,367
RIGHT TO BE INFORMED OF THE 

256
00:09:44,367 --> 00:09:44,968
RESULTS OF THE TRIAL AT THE 

257
00:09:44,968 --> 00:09:46,469
CONCLUSION OF THE STUDY.  THIS 

258
00:09:46,469 --> 00:09:49,906
IS ALL PACKAGED HERE. 

259
00:09:49,906 --> 00:09:52,108
SO THERE'S A LOT OF DIFFERENT 

260
00:09:52,108 --> 00:09:52,676
ASPECTS THAT GO INTO ALL OF 

261
00:09:52,676 --> 00:09:55,879
THIS. AND AGAIN, THIS IS JUST A 

262
00:09:55,879 --> 00:10:03,687
BROAD OVERVIEW OF ALL THESE 

263
00:10:03,687 --> 00:10:04,287
THINGS THAT WE NEED TO THINK 

264
00:10:04,287 --> 00:10:07,824
ABOUT, EVEN BEFORE WE START TO 

265
00:10:07,824 --> 00:10:12,362
ACTUALLY DESIGN THE TRIAL.

266
00:10:12,362 --> 00:10:16,299
NEXT I WANT TO GET INTO THE KEY 

267
00:10:16,299 --> 00:10:26,843
PRINCIPLES. SO COVERING A LOT. 

268
00:10:27,811 --> 00:10:31,181
THIS IS THE MEAT OF THE LECTURE 

269
00:10:31,181 --> 00:10:31,748
HERE. COVERING HOW WE THINK 

270
00:10:31,748 --> 00:10:32,248
ABOUT MAKING CHOICE OF 

271
00:10:32,248 --> 00:10:33,083
DECISIONS.  WHAT THESE THINGS 

272
00:10:33,083 --> 00:10:36,252
ARE.  LOOKING AT THIS FROM THE 

273
00:10:36,252 --> 00:10:38,755
STATISTICAL LENS. AND AGAIN, 

274
00:10:38,755 --> 00:10:43,793
THIS IS A LOT OF MEAT HERE.  A 

275
00:10:43,793 --> 00:10:44,394
LOT OF DETAIL IS MISSING FROM 

276
00:10:44,394 --> 00:10:45,195
THIS PORTION OF THE 

277
00:10:45,195 --> 00:10:48,832
PRESENTATION.  BUT I WANTED TO 

278
00:10:48,832 --> 00:10:50,066
TRY TO GIVE US A GENERAL SCOPE 

279
00:10:50,066 --> 00:10:53,603
AS TO HOW WE NEED TO THINK 

280
00:10:53,603 --> 00:10:54,237
ABOUT THESE ASPECTS WE DESIGN 

281
00:10:54,237 --> 00:10:56,973
THE TRIAL. 

282
00:10:56,973 --> 00:11:07,450
SO, FIRST IS CHOICE OF CONTROL. 

283
00:11:08,218 --> 00:11:08,752
A LOT OF TIMES, WHEN WE THINK 

284
00:11:08,752 --> 00:11:09,686
ABOUT CLINICAL TRIALS, WITH 

285
00:11:09,686 --> 00:11:10,186
THING ABOUT RANDOMIZED, 

286
00:11:10,186 --> 00:11:17,527
PLACEBO-CONTROLLED TRIALS. AND 

287
00:11:17,527 --> 00:11:18,128
IN FACT THIS IS ONLY ONE OF A 

288
00:11:18,128 --> 00:11:19,662
FEW TYPES OF TROUT THAT WE 

289
00:11:19,662 --> 00:11:24,067
MIGHT SEE. PLACEBO TRIALS, 

290
00:11:24,067 --> 00:11:24,667
USING PLACEBO-CONTROLLED IS A 

291
00:11:24,667 --> 00:11:25,435
SCENARIO WHERE WE HAVE A 

292
00:11:25,435 --> 00:11:27,804
TREATMENT THAT IS MORE AROUND 

293
00:11:27,804 --> 00:11:28,905
INVESTIGATIONAL PRODUCT OR 

294
00:11:28,905 --> 00:11:29,506
DRUG, OR WHATEVER ELSE WE ARE 

295
00:11:29,506 --> 00:11:34,711
INVESTIGATING.  BUT IT IS 

296
00:11:34,711 --> 00:11:35,311
CONSIDERED TO BE INACTIVE. SO 

297
00:11:35,311 --> 00:11:37,981
PLACEBO BY NAME, IS GOING TO BE 

298
00:11:37,981 --> 00:11:38,615
THE MOST COMMON CASES WHERE WE 

299
00:11:38,615 --> 00:11:43,420
ARE IN THIS DRUG FORMAT.  WHERE 

300
00:11:43,420 --> 00:11:44,921
WE HAVE SIMILARLY FORMATTED 

301
00:11:44,921 --> 00:11:50,093
PILLS, OR REGIMENTS.  THIS IS 

302
00:11:50,093 --> 00:11:51,761
THE TREATMENT THAT SOME OF THE 

303
00:11:51,761 --> 00:11:54,898
PEOPLE A RANDOMIZED TO RECEIVE. 

304
00:11:54,898 --> 00:11:55,398
THERE ARE OTHER TYPES OF 

305
00:11:55,398 --> 00:11:57,267
CONTROL GROUPS AS WELL BUT WE 

306
00:11:57,267 --> 00:11:57,801
CAN CHOOSE TO USE FOR OUR 

307
00:11:57,801 --> 00:12:00,804
TRIALS. AND EACH IS GOING TO 

308
00:12:00,804 --> 00:12:03,840
HAVE ITS OWN CONSIDERATIONS, 

309
00:12:03,840 --> 00:12:07,010
AND PROS AND CONS TO WHEN THESE 

310
00:12:07,010 --> 00:12:09,045
COULD BE INDICATED FOR USE. 

311
00:12:09,045 --> 00:12:14,751
WE CAN ALSO HAVE A CONTROL ARM 

312
00:12:14,751 --> 00:12:15,251
THAT IS NO TREATMENT OR 

313
00:12:15,251 --> 00:12:15,852
STANDARD OF CARE. MANY OF YOU 

314
00:12:15,852 --> 00:12:17,787
HAVE SEEN TRIALS LIKE THIS AS 

315
00:12:17,787 --> 00:12:18,855
WELL. WHERE PERHAPS WE ARE 

316
00:12:18,855 --> 00:12:23,593
INVESTIGATING LIKE AN EXERCISE 

317
00:12:23,593 --> 00:12:24,794
INTERVENTION, FOR EXAMPLE. IT 

318
00:12:24,794 --> 00:12:26,696
WOULD OBVIOUSLY BE PRETTY HARD 

319
00:12:26,696 --> 00:12:27,297
TO COME UP WITH A PLACEBO FOR 

320
00:12:27,297 --> 00:12:30,667
THAT.  LIKE SOME KIND OF OF THE 

321
00:12:30,667 --> 00:12:32,068
TREATMENT THAT IS NOT THE 

322
00:12:32,068 --> 00:12:34,370
EXERCISE REGIMEN THAT YOU ARE 

323
00:12:34,370 --> 00:12:37,674
INVESTING FOR YOUR PATIENT. 

324
00:12:37,674 --> 00:12:38,208
SO WE WOULD OFTEN TIMES IN 

325
00:12:38,208 --> 00:12:39,242
THOSE SCENARIOS USE KIND OF THE 

326
00:12:39,242 --> 00:12:43,179
STANDARD OF CARE.  MAYBE THEY 

327
00:12:43,179 --> 00:12:43,813
WOULD KEEP GOING TO THE DOCTOR, 

328
00:12:43,813 --> 00:12:44,781
USING ALL THEIR USUAL DRUGS TO 

329
00:12:44,781 --> 00:12:49,953
MANAGE THEIR SYMPTOMS.  AND THE 

330
00:12:49,953 --> 00:12:50,587
TREATMENT GROUP IS RECEIVING AN 

331
00:12:50,587 --> 00:12:52,021
ADDITIONAL, YOU KNOW, EXERCISE  

332
00:12:52,021 --> 00:12:57,327
REGIMEN FOR EXAMPLE.

333
00:12:57,327 --> 00:12:57,894
IN A LOT OF CASES THERE IS NO 

334
00:12:57,894 --> 00:12:59,395
TREATMENT OR STANDARD OF CARE. 

335
00:12:59,395 --> 00:13:00,029
YOU ARE NOT ABLE TO BE BLINDED 

336
00:13:00,029 --> 00:13:00,563
BECAUSE YOU'RE BASICALLY 

337
00:13:00,563 --> 00:13:01,931
TELLING THE PERSON YOU'RE GOING 

338
00:13:01,931 --> 00:13:02,499
TO BE DOING THE SAME THINGS 

339
00:13:02,499 --> 00:13:04,834
YOU'RE GOING TO BE DOING.  OR 

340
00:13:04,834 --> 00:13:07,570
WE ARE NOT PROVIDING YOU AN 

341
00:13:07,570 --> 00:13:10,740
ADDITIONAL TREATMENT. SO WE 

342
00:13:10,740 --> 00:13:11,374
WILL TALK MORE ABOUT BLINDING 

343
00:13:11,374 --> 00:13:14,210
AS WELL.  THIS IS ONE DRAWBACK 

344
00:13:14,210 --> 00:13:14,777
OF THIS PARTICULAR TYPE OF 

345
00:13:14,777 --> 00:13:17,046
CHOICE OF CONTROL. 

346
00:13:17,046 --> 00:13:19,582
WE CAN ALSO HAVE ACTIVE CONTROL.

347
00:13:19,582 --> 00:13:23,853
 SO THIS IS MOST COMMON IN DRUG 

348
00:13:23,853 --> 00:13:24,454
TRIALS, SPECIFICALLY WHERE WE 

349
00:13:24,454 --> 00:13:29,692
ARE ASKING QUESTIONS ABOUT BIO 

350
00:13:29,692 --> 00:13:31,661
EQUIVALENCE OR EQUIVALENT IN 

351
00:13:31,661 --> 00:13:32,295
GENERAL, OR OTHER SIMILAR TYPES 

352
00:13:32,295 --> 00:13:35,932
OF INVESTIGATIONS WERE MAYBE WE 

353
00:13:35,932 --> 00:13:36,466
ARE TRYING TO COMPARE TWO 

354
00:13:36,466 --> 00:13:37,467
TREATMENTS TOGETHER. MAYBE THEY 

355
00:13:37,467 --> 00:13:43,573
ARE RECEIVING -- THEY ARE BEING 

356
00:13:43,573 --> 00:13:44,574
RANDOMIZED TO TREATMENT, BUT IT 

357
00:13:44,574 --> 00:13:45,141
IS NOT GOING TO BE THE ONE 

358
00:13:45,141 --> 00:13:47,343
UNDER INVESTIGATION.  THIS IS 

359
00:13:47,343 --> 00:13:47,944
WHAT WE WOULD CALL AN ACTIVE 

360
00:13:47,944 --> 00:13:51,147
CONTROL. 

361
00:13:51,147 --> 00:13:51,681
WE CAN ALSO HAVE HISTORICAL 

362
00:13:51,681 --> 00:13:54,918
CONTROLS. SO, THIS IS SOMETHING 

363
00:13:54,918 --> 00:13:56,319
I THINK THAT COMES UP A LITTLE 

364
00:13:56,319 --> 00:14:02,559
BIT FOR US AT NINDS.  THIS IS A 

365
00:14:02,559 --> 00:14:06,930
SCENARIO WHERE MAYBE WE HAVE A 

366
00:14:06,930 --> 00:14:07,497
RARE DISEASE, AND MAYBE WE 

367
00:14:07,497 --> 00:14:08,031
DON'T HAVE A WHOLE LOT OF 

368
00:14:08,031 --> 00:14:08,698
PATIENTS TO PULL FROM. MAYBE IT 

369
00:14:08,698 --> 00:14:11,367
IS DIFFICULT TO GET PEOPLE TO 

370
00:14:11,367 --> 00:14:14,370
ENROLL IN TRIALS, WHERE THEY 

371
00:14:14,370 --> 00:14:16,906
MUST BE RANDOMIZED AND SO IT IS 

372
00:14:16,906 --> 00:14:17,473
DIFFICULT TO HAVE A CONTROL 

373
00:14:17,473 --> 00:14:18,608
GROUP AS PART OF THE TRIAL. SO 

374
00:14:18,608 --> 00:14:20,777
THERE ARE SOME SCENARIOS WHERE 

375
00:14:20,777 --> 00:14:27,417
WE HAVE HISTORICAL CONTROL. IT 

376
00:14:27,417 --> 00:14:27,951
IS GOING TO BE A GROUP OF 

377
00:14:27,951 --> 00:14:28,551
PATIENTS WITH THE DISEASE OF 

378
00:14:28,551 --> 00:14:31,521
INTEREST FROM AN EARLIER TIME 

379
00:14:31,521 --> 00:14:32,021
PERIOD MAY BE THESE ARE 

380
00:14:32,021 --> 00:14:33,089
CROSS-SECTIONAL OBSERVATIONAL 

381
00:14:33,089 --> 00:14:35,792
COHORTS. MAY BE NATURAL HISTORY 

382
00:14:35,792 --> 00:14:36,326
STUDY, WHERE WE HAVE SOME 

383
00:14:36,326 --> 00:14:36,960
INFORMATION ON THIS PEOPLE OVER 

384
00:14:36,960 --> 00:14:39,462
TIME.  THIS IS BASICALLY GOING 

385
00:14:39,462 --> 00:14:42,198
TO BE ASYNCHRONOUS. SO THE 

386
00:14:42,198 --> 00:14:42,799
PEOPLE WERE INVESTIGATED AT A 

387
00:14:42,799 --> 00:14:43,766
TIME OUTSIDE OF THE TRIAL THAT 

388
00:14:43,766 --> 00:14:46,436
YOU ARE DESIGNING. 

389
00:14:46,436 --> 00:14:48,538
AND THEN FINALLY, WE CAN HAVE 

390
00:14:48,538 --> 00:14:50,707
MULTIPLE CONTROL GROUPS.  YOU 

391
00:14:50,707 --> 00:14:52,008
CAN USE MORE ONE OF THE ABOVE 

392
00:14:52,008 --> 00:14:53,810
IN ORDER TO DESIGN YOUR TRIAL 

393
00:14:53,810 --> 00:14:57,080
AS WELL.  YOU CAN HAVE BOTH A 

394
00:14:57,080 --> 00:14:59,616
STANDARD CARE, AN ACTIVE 

395
00:14:59,616 --> 00:15:06,589
CONTROL OR A VARIETY OF 

396
00:15:06,589 --> 00:15:07,223
DIFFERENT COMBINATIONS OF THESE 

397
00:15:07,223 --> 00:15:07,857
AS WELL. SO YOU ARE NOT ALWAYS 

398
00:15:07,857 --> 00:15:08,391
LIMITED TO ONE CHOICE OF 

399
00:15:08,391 --> 00:15:16,566
CONTROL.  IT WILL AFFECT YOUR 

400
00:15:16,566 --> 00:15:17,166
REQUIRED SAMPLE SIZE, QUITE A 

401
00:15:17,166 --> 00:15:17,800
BIT AND THEY MIGHT INCREASE THE 

402
00:15:17,800 --> 00:15:25,575
COMPLEXITY OF THE DESIGN. 

403
00:15:25,575 --> 00:15:29,545
CHOICE OF CONTROL IS IMPORTANT 

404
00:15:29,545 --> 00:15:30,847
BECAUSE WE NEED TO  DETERMINE 

405
00:15:30,847 --> 00:15:31,481
HOW MUCH OF THE MEASURED EFFECT 

406
00:15:31,481 --> 00:15:32,081
WE SEE IS DUE TO THE PART OF 

407
00:15:32,081 --> 00:15:32,582
THE STUDY RECEIVING ANY 

408
00:15:32,582 --> 00:15:35,318
PRODUCT.  INCLUDING CONTROL ALL 

409
00:15:35,318 --> 00:15:35,918
RECEIVING THE INVESTIGATIONAL 

410
00:15:35,918 --> 00:15:37,453
TREATMENT. 

411
00:15:37,453 --> 00:15:47,397
SO THIS FIGURE FROM HOWICK ET 

412
00:15:47,397 --> 00:15:48,031
AL IS A GOOD REPRESENTATION OF 

413
00:15:48,031 --> 00:15:48,798
WHY WE USE THESE CONTROLS ARE 

414
00:15:48,798 --> 00:15:49,265
ALL. ILLUSTRATED THE 

415
00:15:49,265 --> 00:15:49,832
DIFFERENCES OF THE TYPES OF 

416
00:15:49,832 --> 00:15:52,969
CONTROL. 

417
00:15:52,969 --> 00:15:53,469
THE FIRST PORTION OF THE 

418
00:15:53,469 --> 00:15:55,505
FIGURE-- I WILL REVEAL THE REST 

419
00:15:55,505 --> 00:15:56,072
OF THE CLICK THROUGH-- THE 

420
00:15:56,072 --> 00:15:56,673
MAGNITUDE OF THE EFFECTS THAT 

421
00:15:56,673 --> 00:15:59,442
WE SEE JUST FROM BEING IN A 

422
00:15:59,442 --> 00:16:01,311
STUDY. EVEN JUST BEING IN A NO 

423
00:16:01,311 --> 00:16:01,878
TREATMENT GROUP. THERE IS A 

424
00:16:01,878 --> 00:16:12,322
NAME FOR THIS. HAWTHORNE 

425
00:16:13,056 --> 00:16:13,623
EFFECT. JUST BEING PART OF THE 

426
00:16:13,623 --> 00:16:14,223
INVESTIGATION HAVING CONTACT 

427
00:16:14,223 --> 00:16:14,757
WITH CLINICIANS, GOING TO 

428
00:16:14,757 --> 00:16:19,162
APPOINTMENTS.  I HESITATE TO 

429
00:16:19,162 --> 00:16:21,464
CALL THIS THE PSYCHOLOGICAL 

430
00:16:21,464 --> 00:16:22,098
EFFECT, BUT IT IS THE EFFECT OF 

431
00:16:22,098 --> 00:16:24,334
BEING OBSERVED AS A PATIENT.  

432
00:16:24,334 --> 00:16:28,905
IN THIS BAR INHERENTLY IS ALSO 

433
00:16:28,905 --> 00:16:37,847
THE TRUE UNDERLYING PROCESSES. 

434
00:16:37,847 --> 00:16:38,448
WHEN AND WHATEVER ADDITIONAL 

435
00:16:38,448 --> 00:16:39,082
FACTORS ARE INCURRED FROM BEING 

436
00:16:39,082 --> 00:16:42,318
OBSERVED.

437
00:16:42,318 --> 00:16:42,885
AND THEN WE HAVE IN THIS CASE 

438
00:16:42,885 --> 00:16:44,320
WHERE WE HAVE A PLACEBO, WE 

439
00:16:44,320 --> 00:16:51,594
HAVE THE ADDITIONAL EFFECT OF 

440
00:16:51,594 --> 00:16:52,228
OBTAINING  A TREATMENT, THAT IS 

441
00:16:52,228 --> 00:16:55,498
THE PLACEBO TREATMENT. THIS IS 

442
00:16:55,498 --> 00:16:56,065
THE PLACEBO EFFECT THAT WE 

443
00:16:56,065 --> 00:16:58,668
TALKED ABOUT.  AGAIN THIS IS 

444
00:16:58,668 --> 00:17:00,169
STILL ON TOP OF THE EFFECT OF 

445
00:17:00,169 --> 00:17:03,740
BEING IN A STUDY AT ALL. 

446
00:17:03,740 --> 00:17:07,844
AND THEN FINALLY IS OUR 

447
00:17:07,844 --> 00:17:09,178
TREATMENT EFFECT HERE AT THE 

448
00:17:09,178 --> 00:17:10,847
TOP, WHICH DEMONSTRATES THE 

449
00:17:10,847 --> 00:17:11,414
ADDITIONAL EFFECT ON TOP OF 

450
00:17:11,414 --> 00:17:14,650
BOTH THE PLACEBO AND THE EFFECT 

451
00:17:14,650 --> 00:17:15,218
OF BEING IN A NO TREATMENT 

452
00:17:15,218 --> 00:17:17,687
GROUP.  SO WE CAN SEE HOW THESE 

453
00:17:17,687 --> 00:17:21,190
THINGS KIND OF STACK. AND I 

454
00:17:21,190 --> 00:17:21,791
THINK STATISTICALLY IT CAN BE 

455
00:17:21,791 --> 00:17:23,359
REALLY IMPORTANT TO MAKE SURE 

456
00:17:23,359 --> 00:17:26,295
THAT YOU ARE OBTAINING AS MUCH 

457
00:17:26,295 --> 00:17:26,896
CONTROL OVER THIS PROCESS AS 

458
00:17:26,896 --> 00:17:30,700
POSSIBLE.  SO YOU ARE ABLE TO 

459
00:17:30,700 --> 00:17:33,269
REALLY ISOLATE AS MUCH AS YOU 

460
00:17:33,269 --> 00:17:33,903
CAN THIS TREATMENT EFFECT THAT 

461
00:17:33,903 --> 00:17:35,838
YOU ARE INTERESTED IN, TO 

462
00:17:35,838 --> 00:17:36,472
REALLY UNDERSTAND THE EFFECT OF 

463
00:17:36,472 --> 00:17:38,374
THE DRUG OR WHATEVER TREATMENT 

464
00:17:38,374 --> 00:17:42,311
YOU ARE INVESTIGATING. 

465
00:17:42,311 --> 00:17:42,912
THE ONLY WAY WE CAN DO THAT IS 

466
00:17:42,912 --> 00:17:53,423
BY THOUGHTFUL CHOICES CONTROL.

467
00:17:54,824 --> 00:17:57,026
NEXT I WANT TO TALK ABOUT 

468
00:17:57,026 --> 00:18:01,731
CHOICE OF ENDPOINTS.  THIS IS 

469
00:18:01,731 --> 00:18:03,433
GOING TO BE PRIMARILY ABOUT THE 

470
00:18:03,433 --> 00:18:06,469
PRIMARY ENDPOINT.  BUT 

471
00:18:06,469 --> 00:18:07,036
OBVIOUSLY WE HAVE A LOT OF 

472
00:18:07,036 --> 00:18:07,637
OTHER ENDPOINTS IN OUR STUDY 

473
00:18:07,637 --> 00:18:12,809
THAT WE NEED TO DEFINED.  SO 

474
00:18:12,809 --> 00:18:14,677
THIS MIGHT BE PERTAINING TO OUR 

475
00:18:14,677 --> 00:18:16,546
PRIMARY ENDPOINT FOR THE STUDY 

476
00:18:16,546 --> 00:18:17,180
AS WELL AS OTHER KEY ENDPOINTS 

477
00:18:17,180 --> 00:18:26,522
THAT WE MIGHT BE INVESTIGATING 

478
00:18:26,522 --> 00:18:27,089
AS PART OF THE TRIAL. THESE 

479
00:18:27,089 --> 00:18:27,723
THINGS ARE GOING TO INFORM THE 

480
00:18:27,723 --> 00:18:28,357
SAMPLE CALCULATION, THE PRIMARY 

481
00:18:28,357 --> 00:18:31,194
ANALYSIS YOU ARE UNDERTAKING, 

482
00:18:31,194 --> 00:18:31,794
GENERALIZABILITY OF WHAT YOU 

483
00:18:31,794 --> 00:18:32,328
FIND IN THE STUDY AND HOW 

484
00:18:32,328 --> 00:18:34,764
APPLICABLE THE STUDY IS.  AND 

485
00:18:34,764 --> 00:18:40,236
HOW USEFUL IT IS IN PRACTICE.

486
00:18:40,236 --> 00:18:40,803
SO THIS CASE OF STUDY DESIGN 

487
00:18:40,803 --> 00:18:41,370
AND TRIAL DESIGN IS REALLY 

488
00:18:41,370 --> 00:18:46,309
IMPORTANT.  AND I THINK WORKING 

489
00:18:46,309 --> 00:18:47,410
WITH THIS AS WELL CAN HELP YOU 

490
00:18:47,410 --> 00:18:54,717
FINE-TUNE THESE THINGS. 

491
00:18:54,717 --> 00:18:55,318
I WANT TO WALK THROUGH SOME OF 

492
00:18:55,318 --> 00:18:55,952
THESE FORMULATIONS OF ENDPOINTS 

493
00:18:55,952 --> 00:18:56,719
LOOK LIKE.  AND HOW YOU MIGHT 

494
00:18:56,719 --> 00:18:58,921
CHOOSE DIFFERENT TYPES OF 

495
00:18:58,921 --> 00:19:02,325
DEFINITIONS OF ENDPOINTS HERE, 

496
00:19:02,325 --> 00:19:02,959
WITH THE UNDERSTANDING THAT OUR 

497
00:19:02,959 --> 00:19:04,927
CHOICE OF THESE THINGS CAN 

498
00:19:04,927 --> 00:19:05,561
REALLY DICTATE THE DIRECTION OF 

499
00:19:05,561 --> 00:19:08,397
OUR TRIAL.  AND THE CHOICE OF 

500
00:19:08,397 --> 00:19:11,434
THESE THINGS SHOULD BE INFORMED 

501
00:19:11,434 --> 00:19:12,869
BY THE PILOT DATA, OR PRIOR 

502
00:19:12,869 --> 00:19:15,338
RESEARCH, OR THE LITERATURE 

503
00:19:15,338 --> 00:19:15,905
HELP YOU FINE-TUNE WHAT IS 

504
00:19:15,905 --> 00:19:19,108
GOING TO BE THE MOST USEFUL 

505
00:19:19,108 --> 00:19:21,677
FORMULATION OF ENDPOINT. WHAT 

506
00:19:21,677 --> 00:19:22,178
IS GOING TO BE THE MOST 

507
00:19:22,178 --> 00:19:27,083
INTERPRETABLE. AND HOW CAN WE 

508
00:19:27,083 --> 00:19:29,385
USE THIS IN THE MOST DIRECTED 

509
00:19:29,385 --> 00:19:33,890
AND COMPLETE WAY?

510
00:19:33,890 --> 00:19:37,460
SO CONTINUES FORMULATION OF 

511
00:19:37,460 --> 00:19:38,027
ENDPOINT, FORMULATION OF A 

512
00:19:38,027 --> 00:19:40,796
METRIC THAT YOU'RE INTERESTED 

513
00:19:40,796 --> 00:19:41,397
IN  FOR YOUR PRIMARY ANALYSIS 

514
00:19:41,397 --> 00:19:41,931
MIGHT BE SOMETHING LIKE A 

515
00:19:41,931 --> 00:19:50,673
DISTANCE WALKED, TIME TO WALK 

516
00:19:50,673 --> 00:19:51,307
EXAMS, ANOTHER COMMON METRIC TO 

517
00:19:51,307 --> 00:19:57,914
USE. PERCENT, VOLUME, AND SOME 

518
00:19:57,914 --> 00:20:00,716
LARGE RANGE SCALE, LIKE STROKE 

519
00:20:00,716 --> 00:20:02,852
SCALE SOMETIMES IS BIG ENOUGH. 

520
00:20:02,852 --> 00:20:03,753
OR IF THERE ARE OTHER LARGER 

521
00:20:03,753 --> 00:20:08,291
RANGE SCALES. SO MEASURES ARE 

522
00:20:08,291 --> 00:20:10,826
GOING TO BE PRIMARILY ON A 

523
00:20:10,826 --> 00:20:12,995
FINE-GRAINED NUMERIC SCALE.  

524
00:20:12,995 --> 00:20:13,529
WHERE YOU CAN THINK ABOUT 

525
00:20:13,529 --> 00:20:16,933
INTUITIVELY TAKING MEANS, OR 

526
00:20:16,933 --> 00:20:19,602
BEING ABLE TO DESCRIBE 

527
00:20:19,602 --> 00:20:20,169
VARIABILITY, AND ALL OF THE 

528
00:20:20,169 --> 00:20:22,438
METRICS YOU MIGHT THINK OF AS 

529
00:20:22,438 --> 00:20:23,005
BEING ABLE TO DO THAT WITH, 

530
00:20:23,005 --> 00:20:26,943
WOULD BE CONSIDERED CONTINUES 

531
00:20:26,943 --> 00:20:28,411
ENDPOINTS. THIS IS EVEN OUTSIDE 

532
00:20:28,411 --> 00:20:32,915
OF THE CHOICE OF WHICH OF THESE 

533
00:20:32,915 --> 00:20:33,449
METRICS SHOULD WE USE TO 

534
00:20:33,449 --> 00:20:35,451
INVESTIGATE? DO WE WANT TO USE 

535
00:20:35,451 --> 00:20:40,923
LIKE A LESION COUNT? OR JUST 

536
00:20:40,923 --> 00:20:44,727
PRESENCE OF LESIONS? SO MANY 

537
00:20:44,727 --> 00:20:45,361
DECISIONS THAT GO INTO CHOOSING 

538
00:20:45,361 --> 00:20:49,432
PRIMARY ENDPOINT OF TRIAL ON 

539
00:20:49,432 --> 00:20:50,032
TOP OF HOW WE MEASURE AND HOW 

540
00:20:50,032 --> 00:20:52,501
WE ARE GOING TO EXTRACT THAT 

541
00:20:52,501 --> 00:20:53,102
INFORMATION. BUT THIS IS JUST 

542
00:20:53,102 --> 00:20:56,606
ONE PIECE OF THE PUZZLE HERE.

543
00:20:56,606 --> 00:21:00,109
BINARY ENDPOINTS, INTUITIVELY 

544
00:21:00,109 --> 00:21:04,347
YES/NO MEASUREMENTS, OR IN SOME 

545
00:21:04,347 --> 00:21:06,582
CASES A BINARIZED CONTINUE 

546
00:21:06,582 --> 00:21:11,621
SCALE SO YOU CAN USE A CUTPOINT 

547
00:21:11,621 --> 00:21:12,254
ON A CONTINUOUS SCALE OR IF YOU 

548
00:21:12,254 --> 00:21:12,888
THINK ABOUT BMI BEING CONTINUE 

549
00:21:12,888 --> 00:21:18,160
SCALE AND BINARIZE ITS WEIGHT 

550
00:21:18,160 --> 00:21:19,428
OR NOT ARE GOING TO BE OF 

551
00:21:19,428 --> 00:21:23,432
INTEREST TO YOU. ALL OF THIS 

552
00:21:23,432 --> 00:21:26,769
WILL BE THINGS LIKE THE 

553
00:21:26,769 --> 00:21:30,006
PRESENCE OR ABSENCE OF LESIONS, 

554
00:21:30,006 --> 00:21:30,573
ALL OF THESE ARE CONSIDERED 

555
00:21:30,573 --> 00:21:38,314
BINARY MEASURES OF AN POINTS. 

556
00:21:38,314 --> 00:21:38,814
A NOMINAL ENDPOINT IS AN 

557
00:21:38,814 --> 00:21:39,415
ENDPOINT WHERE WE HAVE NAMED 

558
00:21:39,415 --> 00:21:43,819
GROUPS.  WHERE WE KNOW INHERENT 

559
00:21:43,819 --> 00:21:45,921
QUANTITATIVE ORDERING. 

560
00:21:45,921 --> 00:21:56,465
SOMETHING LIKE A TUMOR TYPE, OR 

561
00:21:59,435 --> 00:21:59,869
A GENETIC MUTATION TYPE 

562
00:21:59,869 --> 00:22:00,436
ETHNICITY AND NONE OF THESE 

563
00:22:00,436 --> 00:22:01,337
THINGS HAVE A GENERIC ORDER 

564
00:22:01,337 --> 00:22:04,340
TYPE.  YOU CAN HAVE A PRIMARY 

565
00:22:04,340 --> 00:22:06,442
OUTCOME AS NOMINAL FOR EXAMPLE. 

566
00:22:06,442 --> 00:22:06,742
AND

567
00:22:06,742 --> 00:22:08,911
AGAIN ALL OF THESE CHOICES OF 

568
00:22:08,911 --> 00:22:10,479
FORMULATION OF ENDPOINTS ARE 

569
00:22:10,479 --> 00:22:12,381
DIRECTLY RELATED TO THE TYPES 

570
00:22:12,381 --> 00:22:16,419
OF ANALYSIS THAT YOU DO.  THEY 

571
00:22:16,419 --> 00:22:16,986
ARE DIRECTLY RELATED TO THE 

572
00:22:16,986 --> 00:22:17,987
TYPE OF POWER CALCULATIONS THAT 

573
00:22:17,987 --> 00:22:21,290
YOU WOULD DO. AND THE REQUIRED 

574
00:22:21,290 --> 00:22:21,891
SAMPLE SIZE YOU WOULD USE FOR 

575
00:22:21,891 --> 00:22:24,660
YOUR TRIAL.  THIS IS A VERY 

576
00:22:24,660 --> 00:22:26,328
IMPORTANT SET OF DECISIONS THAT 

577
00:22:26,328 --> 00:22:36,706
YOU NEED TO MAKE HERE AS WELL. 

578
00:22:36,706 --> 00:22:37,239
ORDINAL FROM RELATION OF AN 

579
00:22:37,239 --> 00:22:38,374
ENDPOINT ARE NAMED GROUPS THE 

580
00:22:38,374 --> 00:22:42,311
HAVE AN INHERENT ORDER. LIKE 

581
00:22:42,311 --> 00:22:52,121
LABELS MILD/MODERATE/SEVERE. 

582
00:22:52,121 --> 00:22:52,722
EDUCATION LEVELS, DESCRIPTION 

583
00:22:52,722 --> 00:22:53,989
OF THE DATA. PUTTING MARKERS 

584
00:22:53,989 --> 00:22:55,424
ARE GOING TO BE THE MOST COMMON 

585
00:22:55,424 --> 00:23:04,700
TYPES OF FORMULATIONS FOR 

586
00:23:04,700 --> 00:23:05,301
ORDINAL ENDPOINT. AND YOU CAN 

587
00:23:05,301 --> 00:23:07,570
ALSO THINK ABOUT THE LIKERT 

588
00:23:07,570 --> 00:23:08,137
SCALE. NOT A CONTINUE SCALE 

589
00:23:08,137 --> 00:23:09,538
BECAUSE THERE ARE SO FEW 

590
00:23:09,538 --> 00:23:13,943
GROUPS. AND THERE IS THE 

591
00:23:13,943 --> 00:23:14,510
ACKNOWLEDGE GROUP DATA, NOT 

592
00:23:14,510 --> 00:23:15,111
NOMINAL BECAUSE THEY HAVE AN 

593
00:23:15,111 --> 00:23:18,080
INHERENT ORDERING. SOMETHING 

594
00:23:18,080 --> 00:23:18,714
LIKE A LIKERT SCALE WOULD ALSO 

595
00:23:18,714 --> 00:23:19,248
BE CONSIDERED AN ORDINAL 

596
00:23:19,248 --> 00:23:25,321
OUTCOME. 

597
00:23:25,321 --> 00:23:27,056
YOU CAN ALSO HAVE A TRIAL WITH 

598
00:23:27,056 --> 00:23:31,527
THE PRIMARY ENDPOINT OF TIME TO 

599
00:23:31,527 --> 00:23:32,161
EVENT. TIME TO EVENT DATA IS A 

600
00:23:32,161 --> 00:23:32,728
LITTLE TRICKY, IN THE SENSE 

601
00:23:32,728 --> 00:23:34,263
THAT IT IS TWO TYPES OF 

602
00:23:34,263 --> 00:23:37,299
ENDPOINTS TOGETHER. AND SO THE 

603
00:23:37,299 --> 00:23:37,867
FIRST IS A CONTINUOUS TIME 

604
00:23:37,867 --> 00:23:41,971
VARIABLE.  WHICH WOULD BE THE 

605
00:23:41,971 --> 00:23:42,571
TIME ELAPSED FOR OBSERVATION 

606
00:23:42,571 --> 00:23:44,473
FOR THAT PERSON. AND OF BINARY 

607
00:23:44,473 --> 00:23:48,444
EVENT MARKER OF WHETHER OR NOT 

608
00:23:48,444 --> 00:23:49,011
THEY HAVE THE EVENT OR THEY 

609
00:23:49,011 --> 00:23:49,678
WERE CENSORED DURING THE COURSE 

610
00:23:49,678 --> 00:23:53,883
OF THE TRIAL. 

611
00:23:53,883 --> 00:23:54,450
THIS WILL BE LIKE THE TIME TO 

612
00:23:54,450 --> 00:23:55,084
DEATH IN YOUR DIFFERENT GROUPS, 

613
00:23:55,084 --> 00:23:55,651
OR IN YOUR SINGLE GROUP OF 

614
00:23:55,651 --> 00:23:59,221
INTEREST.  TIME TO RELAPSE OR 

615
00:23:59,221 --> 00:24:02,391
TIME TO SYMPTOM RECURRENCE. AND 

616
00:24:02,391 --> 00:24:02,992
SOME TUMOR METRIC PERHAPS AS 

617
00:24:02,992 --> 00:24:09,231
WELL.  THOSE ARE ALL SORTS OF 

618
00:24:09,231 --> 00:24:09,832
TIME TO EVENTS AND ENDPOINTS 

619
00:24:09,832 --> 00:24:15,271
YOU CAN USE. AND AGAIN TIME TO 

620
00:24:15,271 --> 00:24:15,905
EVENT, PRIMARY ENDPOINT WELL IT 

621
00:24:15,905 --> 00:24:18,007
MIGHT BE THE MOST INTERESTING. 

622
00:24:18,007 --> 00:24:20,075
POWERING THE TRIAL, AND THE 

623
00:24:20,075 --> 00:24:21,310
LENGTH OF THE TRIAL YOU WOULD 

624
00:24:21,310 --> 00:24:21,911
NEED TO PROPERLY INVESTIGATE 

625
00:24:21,911 --> 00:24:26,215
THIS.  DO YOU NEED A LONGER 

626
00:24:26,215 --> 00:24:26,816
TRIAL? YOU PROBABLY NEED MORE 

627
00:24:26,816 --> 00:24:27,416
PATIENTS TO DO THIS THE RIGHT 

628
00:24:27,416 --> 00:24:28,484
WAY.  AGAIN THIS IS WHERE ALL 

629
00:24:28,484 --> 00:24:30,519
OF THESE THINGS ARE 

630
00:24:30,519 --> 00:24:35,291
INTERRELATED IN TERMS OF HOW 

631
00:24:35,291 --> 00:24:35,925
YOU PICK ALL OF THESE PIECES TO 

632
00:24:35,925 --> 00:24:41,297
DEVELOP YOUR TRIAL. 

633
00:24:41,297 --> 00:24:43,399
YOU CAN ALSO HAVE COMPOSITE 

634
00:24:43,399 --> 00:24:44,033
ENDPOINTS. WHICH ARE BASICALLY 

635
00:24:44,033 --> 00:24:51,407
MULTIPLE DIFFERENT ENDPOINT.  

636
00:24:51,407 --> 00:24:52,341
USUALLY BINARY, GROUP TOGETHER 

637
00:24:52,341 --> 00:24:55,110
INTO A SINGLE ENDPOINT.  YOU 

638
00:24:55,110 --> 00:24:55,778
CAN HAVE A SINGLE ENDPOINT OF A 

639
00:24:55,778 --> 00:24:57,646
COMBINATION OF PRESENCE OF 

640
00:24:57,646 --> 00:25:00,850
STROKE AND/OR DEATH.  IF A 

641
00:25:00,850 --> 00:25:03,385
PERSON HAS ANY OF THOSE, YOU 

642
00:25:03,385 --> 00:25:08,991
WOULD CALL THAT AN EVENT. IF 

643
00:25:08,991 --> 00:25:09,625
THEY DID NOT HAVE IT YOU WOULD 

644
00:25:09,625 --> 00:25:12,361
HAVE YOUR COMPOSITE OUTCOME. 

645
00:25:12,361 --> 00:25:12,995
AND THIS CAN BE TRICKY AS WELL, 

646
00:25:12,995 --> 00:25:14,163
BECAUSE YOU'RE MAKING A LOT OF 

647
00:25:14,163 --> 00:25:14,763
ASSUMPTIONS ABOUT CLUSTERING 

648
00:25:14,763 --> 00:25:15,564
THESE THINGS TOGETHER.  BUT 

649
00:25:15,564 --> 00:25:16,065
THIS IS ANOTHER TYPE OF 

650
00:25:16,065 --> 00:25:18,701
OUTCOME. 

651
00:25:18,701 --> 00:25:21,303
AND FINALLY, WE HAVE SURROGATE 

652
00:25:21,303 --> 00:25:25,274
ENDPOINT.  USUALLY USED IN 

653
00:25:25,274 --> 00:25:27,409
CASES WHERE WE HAVE A LOT OF 

654
00:25:27,409 --> 00:25:29,278
VERY GOOD PRIOR INFORMATION ON 

655
00:25:29,278 --> 00:25:35,117
THE PROPER PATHWAY OF THE 

656
00:25:35,117 --> 00:25:35,718
SYMPTOMOLOGY AND THE DISEASE 

657
00:25:35,718 --> 00:25:39,455
THAT YOU ARE LOOKING AT.  AND 

658
00:25:39,455 --> 00:25:40,022
WE WOULD HAVE TO BE ABLE TO 

659
00:25:40,022 --> 00:25:41,323
VERY WELL UNDERSTAND THE 

660
00:25:41,323 --> 00:25:42,057
CONNECTEDNESS BETWEEN THE 

661
00:25:42,057 --> 00:25:44,593
OBSERVABLE ENDPOINT, AND THE 

662
00:25:44,593 --> 00:25:47,463
SHORT RANGE ENDPOINT, AND THE 

663
00:25:47,463 --> 00:25:48,697
LONGER-RANGE ENDPOINT THAT 

664
00:25:48,697 --> 00:25:49,265
PERHAPS WE ARE NOT ABLE TO 

665
00:25:49,265 --> 00:25:54,603
OBSERVE.  SURROGATE ENDPOINT 

666
00:25:54,603 --> 00:25:55,237
CAN BE REALLY NICE BECAUSE THEY 

667
00:25:55,237 --> 00:25:55,771
CAN MAKE YOUR TRIAL MORE 

668
00:25:55,771 --> 00:26:00,342
EFFICIENT. BUT YOU HAVE TO HAVE 

669
00:26:00,342 --> 00:26:00,976
REALLY GOOD SCIENTIFIC BACKING 

670
00:26:00,976 --> 00:26:01,944
FOR THE USE OF THESE.

671
00:26:01,944 --> 00:26:07,516
AGAIN YOU ARE JUST USING LIKE 

672
00:26:07,516 --> 00:26:10,486
THE NAME ALLUDES, YOU'RE USING 

673
00:26:10,486 --> 00:26:11,086
ONE METRIC FOR SOMETHING ELSE 

674
00:26:11,086 --> 00:26:12,454
THAT IS PRESUMPTIVE TO BE 

675
00:26:12,454 --> 00:26:15,157
PREDICTIVE OF CLINICAL EVENT.  

676
00:26:15,157 --> 00:26:19,261
IN HIV, WE USE CD4+ COUNT AS A 

677
00:26:19,261 --> 00:26:20,262
SURROGATE FOR DEVELOPMENT OF 

678
00:26:20,262 --> 00:26:24,800
AIDS OR DEATH.  A METRIC WOULD 

679
00:26:24,800 --> 00:26:25,434
USE TO SEE IF OUR TREATMENT IS 

680
00:26:25,434 --> 00:26:28,537
EFFECTIVE.  ASSUMING THAT IF WE 

681
00:26:28,537 --> 00:26:33,909
CAN IMPROVE CD4 COUNT, WE CAN 

682
00:26:33,909 --> 00:26:34,343
IMPROVE THESE OTHER 

683
00:26:34,343 --> 00:26:37,746
LONGER-RANGE OUTCOMES.

684
00:26:37,746 --> 00:26:38,280
SIMILARLY IN CARDIOVASCULAR 

685
00:26:38,280 --> 00:26:40,950
DISEASE OR TRIALS  THAT WE ARE 

686
00:26:40,950 --> 00:26:51,460
INVESTIGATING, THESE SORTS OF 

687
00:26:53,395 --> 00:26:53,929
THINGS AND AFFECTED TREATMENT 

688
00:26:53,929 --> 00:26:54,496
MIGHT HAVE A GOOD IMPACT ON 

689
00:26:54,496 --> 00:26:56,332
BLOOD PRESSURE THAT COULD BE 

690
00:26:56,332 --> 00:26:56,966
INVESTIGATED MUCH FURTHER DOWN 

691
00:26:56,966 --> 00:27:01,003
THE LINE. 

692
00:27:01,003 --> 00:27:06,041
AND THE CHARACTERISTICS OF 

693
00:27:06,041 --> 00:27:06,642
THESE ENDPOINTS SHOULD HAVE, 

694
00:27:06,642 --> 00:27:07,609
REGARDLESS OF WHAT YOU CHOOSE 

695
00:27:07,609 --> 00:27:11,113
HERE, THEY SHOULD ALL BE 

696
00:27:11,113 --> 00:27:17,553
OBJECTIVE.  THEY SHOULD BE-- 

697
00:27:17,553 --> 00:27:18,120
WHEN YOU ARE MEASURING THIS 

698
00:27:18,120 --> 00:27:18,954
THING-- THEY SHOULD BE ABLE TO 

699
00:27:18,954 --> 00:27:19,588
BE OBJECTIVELY MEASURED IN YOUR 

700
00:27:19,588 --> 00:27:21,056
TRIAL.  THE MEASUREMENT SHOULD 

701
00:27:21,056 --> 00:27:22,157
BE REPRODUCIBLE.  YOU SHOULD 

702
00:27:22,157 --> 00:27:27,429
HAVE GOOD INTER AND INTRA 

703
00:27:27,429 --> 00:27:30,199
RELIABILITY. FOR EXAMPLE YOU 

704
00:27:30,199 --> 00:27:34,503
SHOULD UNDERSTAND THE 

705
00:27:34,503 --> 00:27:35,104
VARIABILITY OVER TIME AS WELL 

706
00:27:35,104 --> 00:27:40,376
IN TERMS OF INTERNALLY, OVER 

707
00:27:40,376 --> 00:27:40,909
TIME, ASIDE FROM WHAT IS 

708
00:27:40,909 --> 00:27:42,411
HAPPENING IN THE DISEASE 

709
00:27:42,411 --> 00:27:43,512
PROCESS, IS IT UNBIASED? IS IT 

710
00:27:43,512 --> 00:27:47,616
OBJECTIVE? IS IT GETTING AT 

711
00:27:47,616 --> 00:27:48,250
WHAT WE WANT TO KNOW UNDERLYING 

712
00:27:48,250 --> 00:27:51,687
THE DISEASE? IS THIS CLINICALLY 

713
00:27:51,687 --> 00:27:55,758
RELEVANT?

714
00:27:55,758 --> 00:27:56,358
MAYBE THIS ENDPOINT THAT WE ARE 

715
00:27:56,358 --> 00:27:59,094
INTERESTED IN SOUNDS NICE.  BUT 

716
00:27:59,094 --> 00:27:59,728
MAYBE IT DOESN'T ANSWER WHAT WE 

717
00:27:59,728 --> 00:28:04,533
WANT TO KNOW.  WE NEED TO MAKE 

718
00:28:04,533 --> 00:28:05,501
SURE ALL OF THESE THINGS ARE 

719
00:28:05,501 --> 00:28:06,068
OUTLINED AS WE CHOOSE THESE 

720
00:28:06,068 --> 00:28:07,503
PIECES OF OUR DESIGN. 

721
00:28:07,503 --> 00:28:08,003
IT SHOULD BE FEASIBLE TO 

722
00:28:08,003 --> 00:28:10,706
COLLECT. THOUGH SOMETIMES WHAT 

723
00:28:10,706 --> 00:28:13,375
WE WANT TO KNOW ISN'T WHAT WE 

724
00:28:13,375 --> 00:28:13,976
CAN ACTUALLY COLLECT EASILY, 

725
00:28:13,976 --> 00:28:17,713
WITHIN THE SCALE OF THE TRIAL.  

726
00:28:17,713 --> 00:28:18,313
MAYBE WE REALLY WANT TO DO AN 

727
00:28:18,313 --> 00:28:21,517
EVENT INVESTIGATION.  BUT WE 

728
00:28:21,517 --> 00:28:22,151
DON'T HAVE THE FUNDING, OR THE 

729
00:28:22,151 --> 00:28:29,658
DURATION OF TIME TO CONDUCT THE 

730
00:28:29,658 --> 00:28:32,561
TRIAL THAT WE WANT TO DO WITH 

731
00:28:32,561 --> 00:28:34,263
THE TYPE OF DESIGN AND OUTCOME. 

732
00:28:34,263 --> 00:28:34,897
WE NEED TO THINK ABOUT THAT AS 

733
00:28:34,897 --> 00:28:37,833
WELL. 

734
00:28:37,833 --> 00:28:45,674
THIS MAY SEEM OBVIOUS BUT, WE 

735
00:28:45,674 --> 00:28:46,241
HAVE CHOSEN A PRIORI. THERE 

736
00:28:46,241 --> 00:28:50,846
SHOULD NOT BE REFORMULATING OF 

737
00:28:50,846 --> 00:28:51,480
ENDPOINTS AT SOME POINT IN THE 

738
00:28:51,480 --> 00:28:52,014
TRIAL WHILE WE TALK ABOUT 

739
00:28:52,014 --> 00:28:53,182
ADAPTING DESIGNS LATER. I WILL 

740
00:28:53,182 --> 00:28:55,684
TOUCH ON THAT A LITTLE BIT AS 

741
00:28:55,684 --> 00:28:56,251
WELL. WE CAN CHANGE SAMPLE 

742
00:28:56,251 --> 00:28:58,153
SIZE.  WE CAN CHANGE SOME OF 

743
00:28:58,153 --> 00:29:00,155
THE PIECES OF DESIGN.  WE CAN 

744
00:29:00,155 --> 00:29:00,722
DROP INTERVENTIONS FROM THE 

745
00:29:00,722 --> 00:29:03,292
TRIAL.  BUT CHANGING PRIMARY 

746
00:29:03,292 --> 00:29:07,463
ENDPOINTS IS A LOT TRICKIER.  

747
00:29:07,463 --> 00:29:11,467
BECAUSE THIS REALLY-- YOUR 

748
00:29:11,467 --> 00:29:12,101
PRIMARY ENDPOINT OF INTEREST IS 

749
00:29:12,101 --> 00:29:15,370
DIRECTLY RELATED TO WHETHER OR 

750
00:29:15,370 --> 00:29:15,938
NOT YOU HAVE A POSITIVE OR 

751
00:29:15,938 --> 00:29:18,340
NEGATIVE TRIAL.  SO YOU NEED TO 

752
00:29:18,340 --> 00:29:27,316
MAKE SURE YOU THINK CAREFULLY A 

753
00:29:27,316 --> 00:29:27,916
PRIORI ABOUT THE OUTCOMES AND 

754
00:29:27,916 --> 00:29:30,352
YOUR PRIORITIES. THOSE SHOULD 

755
00:29:30,352 --> 00:29:32,421
NOT CHANGE.

756
00:29:32,421 --> 00:29:35,290
MAYBE IN OUR LAB WE ARE REALLY 

757
00:29:35,290 --> 00:29:35,824
INTERESTED IN LIKE A TIME 

758
00:29:35,824 --> 00:29:40,362
DOESN'T GO GAP. BUT MAYBE OTHER 

759
00:29:40,362 --> 00:29:40,963
INVESTIGATORS IN OTHER PLACES 

760
00:29:40,963 --> 00:29:43,332
DON'T REALLY USE THIS AS MUCH.  

761
00:29:43,332 --> 00:29:43,932
FOR OUR TRIAL WE WANT TO MAKE 

762
00:29:43,932 --> 00:29:48,937
SURE THAT POSITIVE TRIALS WITH 

763
00:29:48,937 --> 00:29:51,707
PRIMARY ENDPOINTS, IT NEEDS TO 

764
00:29:51,707 --> 00:29:55,110
BE SOMETHING THAT OTHER PEOPLE 

765
00:29:55,110 --> 00:29:58,280
WILL FIND CONVINCING. SO, THAT 

766
00:29:58,280 --> 00:29:58,814
PRIMARY ENDPOINT THAT YOU 

767
00:29:58,814 --> 00:30:02,951
CHOOSE FOR YOUR TRIAL SHOULD 

768
00:30:02,951 --> 00:30:03,585
HAVE APPLICABLE IMMEDIATELY TO 

769
00:30:03,585 --> 00:30:06,855
OTHER INSTITUTIONS AND SHOULD 

770
00:30:06,855 --> 00:30:16,899
HOLD ITS OWN IN THE LITERATURE. 

771
00:30:16,899 --> 00:30:18,967
NOW WANT TO TALK ABOUT BLINDING.

772
00:30:18,967 --> 00:30:20,335
 AND OVERALL KIND OF CONTROL OF 

773
00:30:20,335 --> 00:30:25,641
BIAS HERE.  BLINDING IS A 

774
00:30:25,641 --> 00:30:26,175
REALLY POPULAR TOPIC FOR 

775
00:30:26,175 --> 00:30:28,410
CONTROLLED TRIALS.  OBVIOUSLY 

776
00:30:28,410 --> 00:30:29,811
THIS IS ONE OF THE KEY BENEFITS 

777
00:30:29,811 --> 00:30:36,552
OF TRIAL DESIGN. BEING ABLE TO 

778
00:30:36,552 --> 00:30:37,186
BLIND PEOPLE. IT CAN BE REALLY 

779
00:30:37,186 --> 00:30:40,489
IMPORTANT DO THIS. WE WILL TALK 

780
00:30:40,489 --> 00:30:41,089
ABOUT EACH PIECE OF BLINDING 

781
00:30:41,089 --> 00:30:43,058
AND WHY THIS MATTERS.

782
00:30:43,058 --> 00:30:43,625
THERE ARE MULTIPLE PEOPLE IN 

783
00:30:43,625 --> 00:30:44,159
THIS PROCESS THAT CAN BE 

784
00:30:44,159 --> 00:30:44,760
BLINDED OR MASKED. SOMETIMES 

785
00:30:44,760 --> 00:30:47,329
YOU CAN USE THE TERM "MASKING" 

786
00:30:47,329 --> 00:30:55,604
AS WELL. 

787
00:30:55,604 --> 00:30:56,205
SO FIRST YOU CAN HAVE BLINDING 

788
00:30:56,205 --> 00:30:58,574
OF THE PATIENT. HOW WOULD WE 

789
00:30:58,574 --> 00:30:59,174
POSSIBLY BLIND THE PATIENT TO 

790
00:30:59,174 --> 00:30:59,775
WHAT TREATMENT THEY ARE BEING 

791
00:30:59,775 --> 00:31:00,442
RANDOMIZED TO? YOU WANT TO MAKE 

792
00:31:00,442 --> 00:31:01,009
SURE THEY ARE RECEIVING THE 

793
00:31:01,009 --> 00:31:01,944
SAME VISIT SCHEDULES, COMPARED 

794
00:31:01,944 --> 00:31:03,412
TO THE ST. TAMMANY THEY ARE 

795
00:31:03,412 --> 00:31:13,889
RECEIVING DRUGS. WE WANT TO 

796
00:31:21,763 --> 00:31:22,297
MAKE SURE THE DRUGS LOOK THE 

797
00:31:22,297 --> 00:31:22,965
SAME AND ARE SHAPED THE SAME IN 

798
00:31:22,965 --> 00:31:25,601
COLOR.  THIS WOULD BE THE FIRST 

799
00:31:25,601 --> 00:31:26,101
STEP IN BLENDING OF THE 

800
00:31:26,101 --> 00:31:31,907
PATIENT. 

801
00:31:31,907 --> 00:31:32,474
WHY DO WE CARE ABOUT BLINDING 

802
00:31:32,474 --> 00:31:34,443
PATIENTS OR MASKING THE 

803
00:31:34,443 --> 00:31:37,479
PATIENTS? THIS COMES DOWN TO 

804
00:31:37,479 --> 00:31:38,013
PARTICIPANTS THAT NO WHAT 

805
00:31:38,013 --> 00:31:41,917
TREATMENT THEY ARE GETTING 

806
00:31:41,917 --> 00:31:42,551
WOULD BE MORE LIKELY TO CHANGE 

807
00:31:42,551 --> 00:31:45,487
THE BEHAVIOR INTENTIONALLY OR 

808
00:31:45,487 --> 00:31:52,828
UNINTENTIONALLY. AND PERHAPS TO 

809
00:31:52,828 --> 00:31:53,428
SAY NO THEY ARE RECEIVING THE 

810
00:31:53,428 --> 00:31:55,430
CONTROL, OR NOT RECEIVING THE 

811
00:31:55,430 --> 00:31:55,998
TREATMENT THEY WILL BE MORE 

812
00:31:55,998 --> 00:31:56,565
LIKELY TO WITHDRAW FROM THE 

813
00:31:56,565 --> 00:31:57,933
STUDY AND SEEK  TREATMENT 

814
00:31:57,933 --> 00:31:58,533
ELSEWHERE ESPECIALLY IN CASES 

815
00:31:58,533 --> 00:31:59,134
WHERE THEY COULD BE RECEIVING 

816
00:31:59,134 --> 00:32:07,843
OTHER SORTS OF TREATMENT. 

817
00:32:07,843 --> 00:32:08,410
UNINTENTIONALLY THEY MIGHT BE 

818
00:32:08,410 --> 00:32:14,349
MORE OR LESS LIKELY TO REPORT, 

819
00:32:14,349 --> 00:32:14,850
IF THEY THINK THEY ARE 

820
00:32:14,850 --> 00:32:15,350
RECEIVING TREATMENT IS 

821
00:32:15,350 --> 00:32:16,618
BENEFICIAL THEY ARE LIKELY TO 

822
00:32:16,618 --> 00:32:21,223
REPORT POSITIVE SYMPTOMS. IF 

823
00:32:21,223 --> 00:32:21,757
THEY DON'T THINK THEY ARE 

824
00:32:21,757 --> 00:32:22,524
RECEIVING THE ACTIVE TREATMENT 

825
00:32:22,524 --> 00:32:23,091
THAT MIGHT BE MORE DOMESTIC 

826
00:32:23,091 --> 00:32:25,661
ABOUT THE DISEASE COURSE AT THE 

827
00:32:25,661 --> 00:32:29,064
SYMPTOMOLOGY THEY ARE RECEIVING.

828
00:32:29,064 --> 00:32:29,665
 SO IT IS IMPORTANT TO BLIND 

829
00:32:29,665 --> 00:32:36,972
THE PATIENTS. THIS IS SOMETIMES 

830
00:32:36,972 --> 00:32:37,606
THE MOST IMPORTANT ONE, BECAUSE 

831
00:32:37,606 --> 00:32:42,177
THIS IS REALLY THE FIRST STOP. 

832
00:32:42,177 --> 00:32:43,412
YOU CAN ALSO HAVE BLINDING OF 

833
00:32:43,412 --> 00:32:45,447
THE INVESTIGATOR. HOW WE BLIND 

834
00:32:45,447 --> 00:32:46,081
AND INVESTIGATOR IN THE COURSE 

835
00:32:46,081 --> 00:32:46,615
OF THE STUDY? WE DO THIS 

836
00:32:46,615 --> 00:32:47,549
THROUGH RANDOMIZATION. WILL 

837
00:32:47,549 --> 00:32:53,689
TALK ABOUT THAT TOO.  WE DO 

838
00:32:53,689 --> 00:32:54,256
THIS BY CONCEALING THE DRUG 

839
00:32:54,256 --> 00:32:54,756
ASSIGNMENTS ARE CALLING 

840
00:32:54,756 --> 00:32:58,560
TREATMENT A AND B, USING 

841
00:32:58,560 --> 00:33:03,832
ENVELOPES AND LABELED VIALS, 

842
00:33:03,832 --> 00:33:04,466
AND THINGS LIKE THIS. AND WHILE 

843
00:33:04,466 --> 00:33:06,702
WE DO THAT?

844
00:33:06,702 --> 00:33:08,537
INVESTIGATORS WHO KNOW WHAT 

845
00:33:08,537 --> 00:33:10,672
TREATMENT THE INDIVIDUAL 

846
00:33:10,672 --> 00:33:11,206
RECEIVED MIGHT TREAT THE 

847
00:33:11,206 --> 00:33:15,410
PATIENTS DIFFERENTLY. THEY 

848
00:33:15,410 --> 00:33:17,446
MIGHT INSINUATE THAT THEY ARE 

849
00:33:17,446 --> 00:33:19,414
NOT IMPROVING; OR THEY ARE MORE 

850
00:33:19,414 --> 00:33:21,350
BIASED IN THE WAY TO COLLECT 

851
00:33:21,350 --> 00:33:21,950
INFORMATION. WE LIKE TO THINK 

852
00:33:21,950 --> 00:33:26,788
WE ARE NOT DOING THESE THINGS.  

853
00:33:26,788 --> 00:33:27,422
BUT IT IS EASIER TO BE ABLE TO 

854
00:33:27,422 --> 00:33:28,056
BLIND YOURSELF AND NOT HAVE TO 

855
00:33:28,056 --> 00:33:28,623
WORRY ABOUT WHETHER YOU ARE 

856
00:33:28,623 --> 00:33:31,293
DOING A GOOD JOB OF BEING 

857
00:33:31,293 --> 00:33:33,462
UNBIASED OR NOT. SO THIS IS 

858
00:33:33,462 --> 00:33:38,333
ALSO VERY IMPORTANT. 

859
00:33:38,333 --> 00:33:38,867
THIRDLY WE WE CAN ALSO HAVE 

860
00:33:38,867 --> 00:33:40,802
BLINDING OF THE ANALYTICS TEAM, 

861
00:33:40,802 --> 00:33:41,436
OR OTHER INDIVIDUALS INVOLVED 

862
00:33:41,436 --> 00:33:42,904
IN THE ANALYSIS PROCESS, LIKE 

863
00:33:42,904 --> 00:33:45,941
PHYSICIANS, ETC.  I THINK WE 

864
00:33:45,941 --> 00:33:46,575
DON'T THINK ABOUT THIS PART AS 

865
00:33:46,575 --> 00:33:51,246
MUCH. BUT IT IS A TYPE OF 

866
00:33:51,246 --> 00:33:51,780
BLINDING WE CAN HAVE OUR 

867
00:33:51,780 --> 00:33:56,952
TRIALS. 

868
00:33:56,952 --> 00:33:57,386
HOW WOULD WE BLIND AN 

869
00:33:57,386 --> 00:34:01,256
ANALYTICAL TEAM FOR THE TRIAL? 

870
00:34:01,256 --> 00:34:01,857
WE CAN MAINTAIN THE BLINDING 

871
00:34:01,857 --> 00:34:02,491
TREATMENT LABELING TO THE FINAL 

872
00:34:02,491 --> 00:34:12,601
DATA FILE.  AND THERE IS A CODE 

873
00:34:12,601 --> 00:34:13,201
ELSEWHERE AND UNMASK THEM AT 

874
00:34:13,201 --> 00:34:17,806
THE FINAL POINT.  SO THAT IN 

875
00:34:17,806 --> 00:34:18,373
THE COURSE OF ANALYSIS THE 

876
00:34:18,373 --> 00:34:19,007
STATISTICIAN ONLY KNOWS THERE 

877
00:34:19,007 --> 00:34:19,641
ARE TWO GROUPS OR THREE GROUPS, 

878
00:34:19,641 --> 00:34:25,814
ALL LABELED A, B, C, ETC.

879
00:34:25,814 --> 00:34:26,314
THEY VIEW THE ANALYSIS AS 

880
00:34:26,314 --> 00:34:27,549
PRESPECIFIED. AND THEY DON'T 

881
00:34:27,549 --> 00:34:34,189
NEED TO KNOW WHICH GROUP.  YOU 

882
00:34:34,189 --> 00:34:34,790
ARE DOING ALL THE DESCRIPTION 

883
00:34:34,790 --> 00:34:37,292
OF THINGS, EVEN IF YOU KNEW.

884
00:34:37,292 --> 00:34:42,631
AND IN THIS ANALYSIS, IN MOST 

885
00:34:42,631 --> 00:34:43,165
CASES THE STATISTICIAN IS 

886
00:34:43,165 --> 00:34:46,968
SOMEWHAT MORE REMOVED FROM 

887
00:34:46,968 --> 00:34:47,803
BIAS, IN TERMS OF WHAT THEY 

888
00:34:47,803 --> 00:34:50,038
DESIRE TO BE THE OUTCOME THAT 

889
00:34:50,038 --> 00:34:50,672
PERHAPS THE INVESTIGATOR OR THE 

890
00:34:50,672 --> 00:35:01,216
PATIENT. BUT THIS IS NOT ALWAYS 

891
00:35:03,718 --> 00:35:04,319
THE CASE BUT IT IS AN IMPORTANT 

892
00:35:04,319 --> 00:35:04,920
PART OF THE PLANNING PROCESS. 

893
00:35:04,920 --> 00:35:05,554
WE CAN ALSO MEASURE BLINDING OF 

894
00:35:05,554 --> 00:35:07,622
THE PATIENT. AND YOU CAN ASK 

895
00:35:07,622 --> 00:35:08,190
THEM AT THE TRIAL WHAT THEY 

896
00:35:08,190 --> 00:35:11,493
BELIEVED THEY RECEIVED.  YOU 

897
00:35:11,493 --> 00:35:12,127
CAN ASSUME THIS TO BE MORE THAN 

898
00:35:12,127 --> 00:35:15,730
A 50-50, COIN FLIP GUESS. BUT 

899
00:35:15,730 --> 00:35:16,331
YOU CAN SEE IF PEOPLE REALLY 

900
00:35:16,331 --> 00:35:17,199
STARTING TO UNDERSTAND WHAT 

901
00:35:17,199 --> 00:35:21,236
THEY RECEIVED, WHETHER MAYBE 

902
00:35:21,236 --> 00:35:21,736
THE PEOPLE WHO RECEIVED 

903
00:35:21,736 --> 00:35:23,205
TREATMENT HAD A VERY SPECIFIC 

904
00:35:23,205 --> 00:35:33,748
TYPE OF RASH THEY GOT. AND THE 

905
00:35:34,783 --> 00:35:35,283
RECEIVED MORE THIS SORTS OF 

906
00:35:35,283 --> 00:35:35,884
THINGS THAN YOU ANTICIPATED, 

907
00:35:35,884 --> 00:35:38,353
EVEN IF YOU TRY TO BLIND THEM.

908
00:35:38,353 --> 00:35:38,954
THIS IS A SIMPLE DIAGRAM WHICH 

909
00:35:38,954 --> 00:35:40,155
SHOWS HOW WE FLOW THROUGH THE 

910
00:35:40,155 --> 00:35:44,726
BLINDING.  AND WHAT WE CALL THE 

911
00:35:44,726 --> 00:35:51,700
STUDY IN THOSE CASES.

912
00:35:51,700 --> 00:35:52,200
IF THE SUBJECTS WERE NOT 

913
00:35:52,200 --> 00:35:58,440
BLINDED DUE TO TREATMENT 

914
00:35:58,440 --> 00:35:59,040
ASSIGNMENT, WE CALL IT NOT A 

915
00:35:59,040 --> 00:35:59,574
BLINDED STUDY. EVEN IF WE 

916
00:35:59,574 --> 00:36:00,041
MANAGED TO BLIND THE 

917
00:36:00,041 --> 00:36:00,509
INVESTIGATORS OR THE 

918
00:36:00,509 --> 00:36:03,178
STATISTICIANS, WE STILL CALL IT 

919
00:36:03,178 --> 00:36:13,522
AN UNBLINDED STUDY. 

920
00:36:13,755 --> 00:36:14,322
IF WE BLINDED THE SUBJECT AND 

921
00:36:14,322 --> 00:36:15,524
THE INVESTIGATORS, WE CALL IT 

922
00:36:15,524 --> 00:36:16,091
THE DOUBLE-BLIND STUDY. WE 

923
00:36:16,091 --> 00:36:20,629
ADDITIONALLY BLINDED THE 

924
00:36:20,629 --> 00:36:21,229
ANALYTICAL TEAM WE CALL IT A 

925
00:36:21,229 --> 00:36:29,571
TRIPLE BLIND STUDY. YOU ARE 

926
00:36:29,571 --> 00:36:31,973
ABLE TO REPORT IN YOUR 

927
00:36:31,973 --> 00:36:38,847
PROTOCOLS AND FINAL REPORT AND 

928
00:36:38,847 --> 00:36:39,481
PAPER, EXACTLY WHO WAS BLINDED 

929
00:36:39,481 --> 00:36:40,115
AND AT WHAT PHASES. THIS GIVES 

930
00:36:40,115 --> 00:36:45,887
YOU THE TERMINOLOGY FOR THAT.

931
00:36:45,887 --> 00:36:46,354
NOW WANT TO TALK ABOUT 

932
00:36:46,354 --> 00:36:48,690
RANDOMIZATION.  SO THIS IS 

933
00:36:48,690 --> 00:36:51,793
AGAIN ANOTHER REALLY MEATY 

934
00:36:51,793 --> 00:36:55,430
TOPIC THAT IS VERY IMPORTANT.  

935
00:36:55,430 --> 00:36:57,532
BUT THERE'S A LOT OF DETAILS 

936
00:36:57,532 --> 00:36:58,133
HERE IS THAT WE ARE GOING TO 

937
00:36:58,133 --> 00:36:59,935
HAVE TO SKIM OVER, FOR THE 

938
00:36:59,935 --> 00:37:01,236
PURPOSE OF GIVING YOU A GOOD 

939
00:37:01,236 --> 00:37:03,972
OVERALL SCOPE OF TRIAL DESIGN 

940
00:37:03,972 --> 00:37:05,607
HERE. SO THERE'S A HANDFUL OF 

941
00:37:05,607 --> 00:37:07,075
CONTACT THE RANDOMIZATION YOU 

942
00:37:07,075 --> 00:37:07,909
CAN DO FOR YOUR STUDY. 

943
00:37:07,909 --> 00:37:11,613
YOU CAN HAVE A SIMPLE 1:1, THE 

944
00:37:11,613 --> 00:37:14,849
MOST BASIC INVESTIGATION.  WE 

945
00:37:14,849 --> 00:37:16,985
CAN HAVE WEIGHTED RANDOMIZATION 

946
00:37:16,985 --> 00:37:20,589
FOR WE ARE RANDOMIZING MORE 

947
00:37:20,589 --> 00:37:25,994
PEOPLE THAN OTHERS, 1:2, OR 

948
00:37:25,994 --> 00:37:26,628
THREE GROUPS, THREE TREATMENTS; 

949
00:37:26,628 --> 00:37:33,034
WE CAN HAVE A 1:1:1 

950
00:37:33,034 --> 00:37:35,804
RANDOMIZATION. 

951
00:37:35,804 --> 00:37:39,474
WE CAN HAVE LOG RANDOMIZATION 

952
00:37:39,474 --> 00:37:40,375
AND ADAPTIVE RANDOMIZATION. ALL 

953
00:37:40,375 --> 00:37:40,976
OF THE THINGS THAT CAN HAPPEN 

954
00:37:40,976 --> 00:37:43,078
IN THE RANDOMIZATION PROCESS. 

955
00:37:43,078 --> 00:37:44,546
IN GENERAL RANDOMIZATION IS 

956
00:37:44,546 --> 00:37:45,180
CONNECTED BUT HE MUCH LIKE HOW 

957
00:37:45,180 --> 00:37:53,722
YOU WOULD EXPECT.  SO IF IN 

958
00:37:53,722 --> 00:37:54,322
YOUR STUDY, YOU KNOW YOU NEED 

959
00:37:54,322 --> 00:37:55,890
50 PEOPLE FOR YOUR STUDY. AND 

960
00:37:55,890 --> 00:38:00,562
THERE ARE TWO GROUPS IN YOU'RE 

961
00:38:00,562 --> 00:38:01,129
DOING 50-50 RANDOMIZATION, 

962
00:38:01,129 --> 00:38:03,264
ESSENTIALLY WHAT YOUR STEAM 

963
00:38:03,264 --> 00:38:03,798
CONCENTRATION WILL DO OR 

964
00:38:03,798 --> 00:38:05,767
WHOEVER IS TASKED TO GENERATE 

965
00:38:05,767 --> 00:38:08,603
THE RANDOMIZATION LIST IS TO 

966
00:38:08,603 --> 00:38:09,237
GENERATE A PROBABILITY LIST OF 

967
00:38:09,237 --> 00:38:12,073
50% FOR EACH GROUP OF THE 

968
00:38:12,073 --> 00:38:12,641
LENGTH OF YOUR TRIAL SO YOU 

969
00:38:12,641 --> 00:38:13,308
WILL END UP WITH SOMETHING LIKE 

970
00:38:13,308 --> 00:38:23,318
IN AB, AAB, ABABB, ETC, ENDING 

971
00:38:23,318 --> 00:38:25,487
WITH A RANDOMIZED GROUP OR 

972
00:38:25,487 --> 00:38:26,121
GREATER IF YOU WOULD ALLOW FOR 

973
00:38:26,121 --> 00:38:27,455
DROPOUTS AND THINGS LIKE THAT. 

974
00:38:27,455 --> 00:38:27,989
AS YOU RANDOMIZE PEOPLE YOU 

975
00:38:27,989 --> 00:38:34,829
FOLLOW THIS LIST AND ADMINISTER 

976
00:38:34,829 --> 00:38:35,430
THE TREATMENT ACCORDINGLY FOR 

977
00:38:35,430 --> 00:38:37,465
EACH NEW LIEN ENROLLED PATIENT.

978
00:38:37,465 --> 00:38:44,239
THE OTHER TYPES A RANDOMIZATION 

979
00:38:44,239 --> 00:38:44,839
ARE BASICALLY ALTERATIONS OF 

980
00:38:44,839 --> 00:38:45,473
THAT BASIC THEME, REALLY IT IS 

981
00:38:45,473 --> 00:38:47,342
NOT THAT MUCH MORE COMPLICATED. 

982
00:38:47,342 --> 00:38:47,942
 AGAIN YOUR STATISTICIAN CAN 

983
00:38:47,942 --> 00:38:49,444
HELP YOU WITH THE STUFF. 

984
00:38:49,444 --> 00:38:57,786
IN CASES WHERE YOU HAVE THIS 

985
00:38:57,786 --> 00:38:58,420
WEIGHTED   RANDOMIZATION OR IF 

986
00:38:58,420 --> 00:38:58,887
YOU WANT MORE PEOPLE 

987
00:38:58,887 --> 00:38:59,587
PREFERENTIALLY IN YOUR 

988
00:38:59,587 --> 00:39:00,155
TREATMENT GROUP OR IN YOUR 

989
00:39:00,155 --> 00:39:01,890
CONTROL GROUP OR A 2:1 

990
00:39:01,890 --> 00:39:02,457
RANDOMIZATION YOU ARE JUST 

991
00:39:02,457 --> 00:39:03,458
CHANGING THE PROBABILITY HERE. 

992
00:39:03,458 --> 00:39:09,464
YOUR TREATMENT LIST WILL HAVE 

993
00:39:09,464 --> 00:39:12,600
2:1 RANDOMIZATION AND 56% IN 

994
00:39:12,600 --> 00:39:13,234
YOUR TREATMENT GROUP AND 33% IN 

995
00:39:13,234 --> 00:39:15,003
THE CONTROL GROUP. 

996
00:39:15,003 --> 00:39:16,404
IF YOU HAVE THREE GROUPS, NOW 

997
00:39:16,404 --> 00:39:25,080
YOU HAVE 33, 33, 33%. AND AGAIN 

998
00:39:25,080 --> 00:39:25,647
ABC, RANDOMIZING THE WHOLE 

999
00:39:25,647 --> 00:39:28,783
LIST. 

1000
00:39:28,783 --> 00:39:31,553
WHERE IT GETS A TINY BIT MORE 

1001
00:39:31,553 --> 00:39:32,120
COMPLICATED IS WHEN WE TALK 

1002
00:39:32,120 --> 00:39:33,788
ABOUT BLOCK RANDOMIZATION. OR 

1003
00:39:33,788 --> 00:39:43,498
BLOCK BEHAVIOR AND 

1004
00:39:43,498 --> 00:39:44,099
RANDOMIZATION AND STRATIFIED 

1005
00:39:44,099 --> 00:39:50,905
RANDOMIZATION  FOR BLOCK 

1006
00:39:50,905 --> 00:39:51,539
RANDOMIZATION THE SHUFFLING OF 

1007
00:39:51,539 --> 00:39:52,140
GROUP LABELS HAPPENS WITHIN A 

1008
00:39:52,140 --> 00:39:52,640
SPECIFIC BLOCK SIZE OF 

1009
00:39:52,640 --> 00:39:54,809
PATIENTS.  SO AS WE CAN SEE 

1010
00:39:54,809 --> 00:39:57,312
HERE AT THE TOP OF THIS 

1011
00:39:57,312 --> 00:39:57,946
EXAMPLE, IN THE FIRST BLOCK OF 

1012
00:39:57,946 --> 00:39:59,013
FOUR PATIENTS WE HAVE THREE 

1013
00:39:59,013 --> 00:40:00,515
PEOPLE RANDOMIZED FOR A, ONE TO 

1014
00:40:00,515 --> 00:40:05,520
B.  IF WE CARE ABOUT 

1015
00:40:05,520 --> 00:40:09,691
MAINTAINING THE BALANCE OF 

1016
00:40:09,691 --> 00:40:11,693
RANDOMIZED INDIVIDUALS TO EACH 

1017
00:40:11,693 --> 00:40:15,296
TREATMENT, AS THE TRIAL 

1018
00:40:15,296 --> 00:40:15,864
PROCEEDS, IF WE THINK THAT 

1019
00:40:15,864 --> 00:40:16,464
MAYBE THERE'S GOING TO BE AN 

1020
00:40:16,464 --> 00:40:18,199
EARLY STOPPING. OR PERHAPS 

1021
00:40:18,199 --> 00:40:18,833
THERE IS SOME OTHER REASON WHY 

1022
00:40:18,833 --> 00:40:21,970
WE WANT TO MAINTAIN BALANCE,  

1023
00:40:21,970 --> 00:40:25,673
OVER THE COURSE OF THE TRIAL AS 

1024
00:40:25,673 --> 00:40:28,309
YOU RANDOMIZE PEOPLE, THEN WE 

1025
00:40:28,309 --> 00:40:28,710
CAN CONSTRAIN THE 

1026
00:40:28,710 --> 00:40:30,178
RANDOMIZATION.  SUCH THAT 

1027
00:40:30,178 --> 00:40:33,348
WITHIN EACH BLOCK OF FOUR 

1028
00:40:33,348 --> 00:40:34,949
THERE'S IN THIS CASE IF YOU 

1029
00:40:34,949 --> 00:40:37,152
TREATMENT, TWO AND TWO WITHIN 

1030
00:40:37,152 --> 00:40:40,655
EACH BLOCK OF FOUR. 

1031
00:40:40,655 --> 00:40:42,056
SO THE LIST CAN STILL RUN.  WE 

1032
00:40:42,056 --> 00:40:44,392
WOULD STILL ADMINISTER THE 

1033
00:40:44,392 --> 00:40:48,096
FIRST PATIENT A, BBA, THEN WE 

1034
00:40:48,096 --> 00:40:55,603
WOULD ADMINISTER A, BAB, AND SO 

1035
00:40:55,603 --> 00:40:59,841
FORTH.  SO YOU WOULD HAVE THESE 

1036
00:40:59,841 --> 00:41:06,080
CHECK POINTS AT EACH OF THESE 

1037
00:41:06,080 --> 00:41:06,648
ENDS OF THE FOUR BLOCKS TO 

1038
00:41:06,648 --> 00:41:07,182
ENSURE THAT YOU HAVE THIS 

1039
00:41:07,182 --> 00:41:11,319
BALANCE AS YOU TREAT. 

1040
00:41:11,319 --> 00:41:12,587
FOR STRATIFIED RANDOMIZATION, 

1041
00:41:12,587 --> 00:41:14,756
AGAIN THE SAME IDEA OF 

1042
00:41:14,756 --> 00:41:15,857
GENERATING A RANDOMIZATION LIST 

1043
00:41:15,857 --> 00:41:19,527
IS THE SAME CONCEPT.  BUT WE 

1044
00:41:19,527 --> 00:41:20,495
USE STRATIFIED RANDOMIZATION IN 

1045
00:41:20,495 --> 00:41:21,062
CASES WHERE WE THINK THERE 

1046
00:41:21,062 --> 00:41:24,399
MIGHT BE AN INFLUENCE OF SOME 

1047
00:41:24,399 --> 00:41:27,936
KIND OF VARIABLE.  OR WE WANT 

1048
00:41:27,936 --> 00:41:30,004
TO MAKE SURE THERE IS AN EQUAL 

1049
00:41:30,004 --> 00:41:30,605
NUMBER OF PEOPLE WITH CERTAIN 

1050
00:41:30,605 --> 00:41:31,239
CHARACTERISTICS IN EACH OF THE 

1051
00:41:31,239 --> 00:41:32,307
TREATMENT GROUPS. 

1052
00:41:32,307 --> 00:41:33,675
FOR EXAMPLE IN YOUR TRIAL, YOU 

1053
00:41:33,675 --> 00:41:37,111
THINK THAT SEX AND AGE IS 

1054
00:41:37,111 --> 00:41:39,814
HIGHLY RELATED TO HOW A PERSON 

1055
00:41:39,814 --> 00:41:40,448
WILL RESPOND TO TREATMENT, THAT 

1056
00:41:40,448 --> 00:41:42,851
WE CREATE THESE STRATIFICATION 

1057
00:41:42,851 --> 00:41:45,720
LEVELS.  WHERE WE WOULD THEN 

1058
00:41:45,720 --> 00:41:55,730
BREAK APART INDIVIDUALS BY SEX 

1059
00:41:55,730 --> 00:41:59,067
AND AGE, AND WE GENERATE A 

1060
00:41:59,067 --> 00:41:59,634
RANDOMIZATION LIST FOR EACH 

1061
00:41:59,634 --> 00:42:05,673
GROUP AND FOR EXAMPLE MALES, 

1062
00:42:05,673 --> 00:42:09,644
>50 YEARS OLD, AND THEN FOR 

1063
00:42:09,644 --> 00:42:10,278
YOUNG MALES WE HAVE THE BALANCE 

1064
00:42:10,278 --> 00:42:10,879
OF STRATIFICATION BALANCE IN 

1065
00:42:10,879 --> 00:42:12,380
EACH OF THE STRATA.  AND 

1066
00:42:12,380 --> 00:42:12,981
HONESTLY WOULD HAVE TWO OTHER 

1067
00:42:12,981 --> 00:42:15,650
STRATIFICATION LEVELS, COULD BE 

1068
00:42:15,650 --> 00:42:16,251
FEMALE, YOUNG AND OLD IN THIS 

1069
00:42:16,251 --> 00:42:17,452
EXAMPLE.

1070
00:42:17,452 --> 00:42:23,124
AND AGAIN THIS ENSURES THAT WE 

1071
00:42:23,124 --> 00:42:24,592
ARE MAINTAINING AN EQUAL NUMBER 

1072
00:42:24,592 --> 00:42:27,128
OF TREATED PEOPLE WITH EACH OF 

1073
00:42:27,128 --> 00:42:31,165
THE LEVELS OF INTEREST. 

1074
00:42:31,165 --> 00:42:32,734
AND AGAIN, WE CAN HAVE ADAPTIVE 

1075
00:42:32,734 --> 00:42:35,737
RANDOMIZATION.  I AM NOT GOING 

1076
00:42:35,737 --> 00:42:36,337
TO GO INTO A LOT OF DETAILS 

1077
00:42:36,337 --> 00:42:42,010
HERE.  BUT WE CAN RANDOMIZE 

1078
00:42:42,010 --> 00:42:42,610
PEOPLE TO TREATMENTS THAT ARE 

1079
00:42:42,610 --> 00:42:43,611
DOING BETTER IN THE COURSE OF 

1080
00:42:43,611 --> 00:42:45,713
THE TRIAL OR ALTER THE 

1081
00:42:45,713 --> 00:42:48,116
RANDOMIZATION DURING THE TRIAL 

1082
00:42:48,116 --> 00:42:50,685
BASED ON SOME TREATMENT DESIGN. 

1083
00:42:50,685 --> 00:42:51,119
 IT SHOULD STILL BE 

1084
00:42:51,119 --> 00:43:01,362
PRESPECIFIED. 

1085
00:43:01,596 --> 00:43:02,196
AND I WANT TO STEP THROUGH SOME 

1086
00:43:02,196 --> 00:43:04,065
COMMON DESIGNS, JUST TO GIVE 

1087
00:43:04,065 --> 00:43:06,067
YOU A SENSE OF THE DIFFERENT 

1088
00:43:06,067 --> 00:43:06,701
FLAVORS OF CLINICAL TRIALS THAT 

1089
00:43:06,701 --> 00:43:10,605
YOU CAN HAVE. AND AGAIN, THE 

1090
00:43:10,605 --> 00:43:11,105
SCAN TO BE A VERY LARGE 

1091
00:43:11,105 --> 00:43:13,341
DECISION TO MAKE IN YOUR DESIGN 

1092
00:43:13,341 --> 00:43:19,681
PROCESS. AND I AM TRYING TO 

1093
00:43:19,681 --> 00:43:20,281
JUST PRESENT THIS IN A SIMPLE 

1094
00:43:20,281 --> 00:43:21,149
WAY, SOME OF THESE OBVIOUSLY 

1095
00:43:21,149 --> 00:43:23,518
FOR PAROL DESIGNS FOR EXAMPLE 

1096
00:43:23,518 --> 00:43:24,252
YOU CAN HAVE MORE THAN TWO ARMS 

1097
00:43:24,252 --> 00:43:31,159
HERE, ETC. 

1098
00:43:31,159 --> 00:43:31,759
SO FOR APPAREL DESIGN, YOU HAVE 

1099
00:43:31,759 --> 00:43:32,694
YOUR INDIVIDUALS, YOU RANDOMIZE 

1100
00:43:32,694 --> 00:43:36,364
TO ONE OF TWO OR MORE GROUPS.  

1101
00:43:36,364 --> 00:43:36,798
AND YOU FOLLOW THEM 

1102
00:43:36,798 --> 00:43:40,935
CONCURRENTLY OVER TIME. AND 

1103
00:43:40,935 --> 00:43:43,504
AGAIN, THE MAIN GIST OF THIS 

1104
00:43:43,504 --> 00:43:44,138
HERE IS THAT THEY ARE RECEIVING 

1105
00:43:44,138 --> 00:43:47,208
TREATMENT AT THE SAME TIME 

1106
00:43:47,208 --> 00:43:50,044
POINT AND UNDER THAT TREATMENT 

1107
00:43:50,044 --> 00:43:51,312
COURSE, THERE MAY BE CERTAIN 

1108
00:43:51,312 --> 00:43:51,913
MEASUREMENTS FOLLOWING OR NOT 

1109
00:43:51,913 --> 00:43:54,082
FOLLOWING THE SAME ONE EITHER 

1110
00:43:54,082 --> 00:43:57,352
ACTIVE TREATMENT OR PLACEBO.  

1111
00:43:57,352 --> 00:43:57,952
WHATEVER YOU OTHER CHOICE OF 

1112
00:43:57,952 --> 00:44:01,789
CONTROL IS, NOTED SIMPLY BY A 

1113
00:44:01,789 --> 00:44:03,691
MEDICINE ALTHOUGH THIS CAN BE 

1114
00:44:03,691 --> 00:44:04,659
WHATEVER THE INVESTIGATIONAL 

1115
00:44:04,659 --> 00:44:07,328
PRODUCT YOU HAVE. 

1116
00:44:07,328 --> 00:44:07,895
THIS IS GOING TO BE LIKE THE 

1117
00:44:07,895 --> 00:44:14,969
MOST SIMPLE TYPE OF DESIGN YOU 

1118
00:44:14,969 --> 00:44:15,603
CAN THINK OF WHEN TALKING ABOUT 

1119
00:44:15,603 --> 00:44:17,138
RANDOMIZED CONTROLLED TRIAL

1120
00:44:17,138 --> 00:44:18,740
YOU CAN ALSO HAVE A CROSSOVER 

1121
00:44:18,740 --> 00:44:19,273
DESIGN WHERE PEOPLE ARE 

1122
00:44:19,273 --> 00:44:20,708
ESSENTIALLY RANDOMIZED TO AN 

1123
00:44:20,708 --> 00:44:23,644
ORDER OF ALL DRUGS USUALLY IN 

1124
00:44:23,644 --> 00:44:27,615
THE TRIAL. SO THEY MIGHT BE 

1125
00:44:27,615 --> 00:44:29,117
RANDOMIZED TO RECEIVE TREATMENT 

1126
00:44:29,117 --> 00:44:31,419
A FIRST AND THEN TREATMENT B. 

1127
00:44:31,419 --> 00:44:33,121
OR RANDOMIZED TO RECEIVE 

1128
00:44:33,121 --> 00:44:33,988
TREATMENT B FIRST AND THEN 

1129
00:44:33,988 --> 00:44:37,325
TREATMENT A.  THERE ARE A LOT 

1130
00:44:37,325 --> 00:44:37,959
OF CONSIDERATIONS FOR CROSSOVER 

1131
00:44:37,959 --> 00:44:42,730
DESIGNS.  WASHOUT PERIODS.  HOW 

1132
00:44:42,730 --> 00:44:43,164
LONG WE WAIT BEFORE 

1133
00:44:43,164 --> 00:44:46,968
ADMINISTERING THE SECOND DRUG. 

1134
00:44:46,968 --> 00:44:49,404
IS THERE A CARRYOVER EFFECT 

1135
00:44:49,404 --> 00:44:50,004
THAT WE NEED TO BE CONCERNED 

1136
00:44:50,004 --> 00:45:00,515
ABOUT? DROPOUTS. FOLLOWING UP 

1137
00:45:03,885 --> 00:45:04,452
WITH PEOPLE FOR THIS TO BE TOO 

1138
00:45:04,452 --> 00:45:07,455
LONG. IN THE SECOND PORTION OF 

1139
00:45:07,455 --> 00:45:08,456
THE CROSSOVER. THIS IS NOT 

1140
00:45:08,456 --> 00:45:09,090
ANOTHER UNCOMMON DESIGN I WOULD 

1141
00:45:09,090 --> 00:45:15,663
SAY, OF CROSSOVER TRIAL.

1142
00:45:15,663 --> 00:45:18,299
IF DONE PROPERLY IF RANDOMIZED 

1143
00:45:18,299 --> 00:45:21,269
WITH SUBJECT DATA, TO MAKING 

1144
00:45:21,269 --> 00:45:21,869
SURE THERE IS NO CARRYOVER OR 

1145
00:45:21,869 --> 00:45:23,337
WASHOUT DURING TREATMENT.

1146
00:45:23,337 --> 00:45:31,079
ALSO YOU HAVE N-OF-1 STUDIES, 

1147
00:45:31,079 --> 00:45:31,679
RELATIVELY UNHEARD OF. MAYBE 

1148
00:45:31,679 --> 00:45:34,048
YOU HAVE A VERY RARE, ULTRA 

1149
00:45:34,048 --> 00:45:37,185
RARE DISEASE.  IT CAN BE 

1150
00:45:37,185 --> 00:45:37,752
DIFFICULT TO DESIGN A FULL 

1151
00:45:37,752 --> 00:45:39,654
TRIAL. OR THERE IS SOME 

1152
00:45:39,654 --> 00:45:42,223
SPECIFIC REASON WHY YOU WANT TO 

1153
00:45:42,223 --> 00:45:43,257
LOOK AT A SPECIFIC PATIENT. YOU 

1154
00:45:43,257 --> 00:45:48,129
COULD RANDOMIZE, YOU COULD 

1155
00:45:48,129 --> 00:45:48,763
FOLLOW ONE INDIVIDUAL UNTREATED 

1156
00:45:48,763 --> 00:45:51,032
AND RANDOMIZE THEM TO ONE OR 

1157
00:45:51,032 --> 00:45:56,204
MORE TREATMENT THE NOTED HERE 

1158
00:45:56,204 --> 00:45:56,804
BUT YOU HAVE SOME FLEXIBILITY 

1159
00:45:56,804 --> 00:45:57,371
IN THE DESIGN TO DO A DEEP 

1160
00:45:57,371 --> 00:45:57,972
INVESTIGATION ON ONE SPECIFIC 

1161
00:45:57,972 --> 00:45:59,474
PERSON.  THERE'S A LOT OF 

1162
00:45:59,474 --> 00:46:00,041
LITERATURE ON THE SPECIFIC 

1163
00:46:00,041 --> 00:46:04,912
DESIGN.  BUT THIS IS ANOTHER 

1164
00:46:04,912 --> 00:46:10,384
TYPE OF DESIGN THAT CAN COME UP.

1165
00:46:10,384 --> 00:46:12,320
WE ALSO HAVE STEPPED WEDGE 

1166
00:46:12,320 --> 00:46:14,388
DESIGNS, WHERE PEOPLE 

1167
00:46:14,388 --> 00:46:15,690
RANDOMIZED TO THE TIME OF 

1168
00:46:15,690 --> 00:46:19,127
STAGGER OF THE DRUG. IN THIS 

1169
00:46:19,127 --> 00:46:20,628
CASE EVERYBODY WILL RECEIVE THE 

1170
00:46:20,628 --> 00:46:21,162
DRUG, OR THE TREATMENT OF 

1171
00:46:21,162 --> 00:46:22,930
INTEREST, WHICH IS GOOD FOR 

1172
00:46:22,930 --> 00:46:24,599
SOME MEDICAL REASONS OR SOME 

1173
00:46:24,599 --> 00:46:30,204
LOGISTICAL REASONS.  MAYBE WE 

1174
00:46:30,204 --> 00:46:30,771
CAN'T ROLL OUR TREATMENT TO 

1175
00:46:30,771 --> 00:46:32,773
EVERYBODY AT THE SAME TIME.

1176
00:46:32,773 --> 00:46:33,374
SO EVERYBODY IS ABSURD FOR SOME 

1177
00:46:33,374 --> 00:46:34,976
BASELINE AMOUNT OF TIME. AT 

1178
00:46:34,976 --> 00:46:37,478
LEAST SOME AMOUNT OF TIME, AND 

1179
00:46:37,478 --> 00:46:38,446
THEN THEY ARE ROLLED INTO 

1180
00:46:38,446 --> 00:46:43,017
TREATMENT OVER TIME AND THEN 

1181
00:46:43,017 --> 00:46:43,618
FOLLOWED FOR THE REST OF THE 

1182
00:46:43,618 --> 00:46:44,752
TRIAL.

1183
00:46:44,752 --> 00:46:50,024
WE CAN ALSO HAVE FACTORIAL 

1184
00:46:50,024 --> 00:46:51,492
DESIGNS, RANDOMIZED TO 

1185
00:46:51,492 --> 00:46:55,429
COMBINATIONS OF TREATMENT 

1186
00:46:55,429 --> 00:46:56,264
CONCURRENTLY. MAYBE BECAUSE THE 

1187
00:46:56,264 --> 00:46:56,831
CASE WE ARE INVESTIGATING, 

1188
00:46:56,831 --> 00:46:59,200
PAIRS OF PLACEBOS -- THIS IS 

1189
00:46:59,200 --> 00:47:00,568
ANOTHER NOT UNCOMMON TYPE OF 

1190
00:47:00,568 --> 00:47:06,941
DESIGN.  THIS CAN BE VERY 

1191
00:47:06,941 --> 00:47:07,542
EFFICIENT IF YOU'RE TRYING TO 

1192
00:47:07,542 --> 00:47:08,042
LEARN THAN ONE DRUG FOR 

1193
00:47:08,042 --> 00:47:15,750
TREATMENT. 

1194
00:47:15,750 --> 00:47:16,284
WE ALSO HAVE MATCHED PAIRS 

1195
00:47:16,284 --> 00:47:17,185
DESIGN. IF WE HAVE VERY GOOD 

1196
00:47:17,185 --> 00:47:18,619
EVIDENCE OR VERY SPECIFIC 

1197
00:47:18,619 --> 00:47:19,954
FRAMEWORK WHERE WE HAVE NATURAL 

1198
00:47:19,954 --> 00:47:22,523
PAIRING IN OUR DATA, WE CAN 

1199
00:47:22,523 --> 00:47:23,824
RANDOMIZE WITHIN PAIRS OF 

1200
00:47:23,824 --> 00:47:27,328
PEOPLE TO TREATMENT OR CONTROL. 

1201
00:47:27,328 --> 00:47:28,896
AND THAT ANALYTICALLY, WE CAN 

1202
00:47:28,896 --> 00:47:30,698
COMPARE BEHAVIOR WITHIN THESE 

1203
00:47:30,698 --> 00:47:35,603
PAIRS.  THE MOST OF HIS EXAMPLE 

1204
00:47:35,603 --> 00:47:36,237
WILL BE LIKE A JOINT STUDY FOR 

1205
00:47:36,237 --> 00:47:39,140
EXAMPLE WHAT WE HAVE BUILT IN 

1206
00:47:39,140 --> 00:47:39,674
PAIRS, OR OTHER TYPES OF 

1207
00:47:39,674 --> 00:47:40,274
GENETIC STUDIES WHERE YOU CAN 

1208
00:47:40,274 --> 00:47:42,977
RANDOMIZE THOSE PAIRS OF PEOPLE 

1209
00:47:42,977 --> 00:47:43,611
WITH A MORE GENETIC PROFILE TO 

1210
00:47:43,611 --> 00:47:44,178
GENETIC TREATMENTS. ANOTHER 

1211
00:47:44,178 --> 00:47:48,916
FRAMEWORK. 

1212
00:47:48,916 --> 00:47:51,285
WE CAN HAVE MORE ADVANCED 

1213
00:47:51,285 --> 00:47:54,956
DESIGNS. GOING INTO THE WEEDS 

1214
00:47:54,956 --> 00:47:58,859
HERE A LITTLE BIT. SEAMLESS II 

1215
00:47:58,859 --> 00:48:01,729
OR III DESIGN WHERE WE CAN DROP 

1216
00:48:01,729 --> 00:48:06,867
TREATMENT AFTER SOME PERIOD OF 

1217
00:48:06,867 --> 00:48:07,468
TIME. MOVING INTO A FOUR LOOK 

1218
00:48:07,468 --> 00:48:08,569
AT PHASE 3 EFFICACY WE ARE 

1219
00:48:08,569 --> 00:48:09,170
DETERMINING FROM THE ANALYSIS 

1220
00:48:09,170 --> 00:48:09,704
THAT THERE IS A SPECIFIC 

1221
00:48:09,704 --> 00:48:14,242
TREATMENT THAT IS NOT BEING AS 

1222
00:48:14,242 --> 00:48:14,842
BENEFICIAL WE CAN ROLL INTO A 

1223
00:48:14,842 --> 00:48:15,409
MORE RIGOROUS LOOK AT FEWER 

1224
00:48:15,409 --> 00:48:20,081
DRUGS. 

1225
00:48:20,081 --> 00:48:22,350
AND OTHER TYPES OF ADAPTIVE 

1226
00:48:22,350 --> 00:48:23,985
DESIGNS, INCLUDING INTERNAL 

1227
00:48:23,985 --> 00:48:26,887
PILOTING. IT IS KIND OF 

1228
00:48:26,887 --> 00:48:30,157
ADAPTIVE, THE SAME FLAVOR.  

1229
00:48:30,157 --> 00:48:35,096
AFTER SOME PERIOD OF TIME WE 

1230
00:48:35,096 --> 00:48:35,730
MIGHT DO SOME KIND OF RANDOMIZE 

1231
00:48:35,730 --> 00:48:39,767
PROCESS OR OTHER WAYS OF 

1232
00:48:39,767 --> 00:48:40,368
COLLECTING PATIENTS, WHETHER 

1233
00:48:40,368 --> 00:48:45,373
PART OF THE PILOT OR ROLLING 

1234
00:48:45,373 --> 00:48:48,909
INTO A STUDY. AND MAKING SURE 

1235
00:48:48,909 --> 00:48:49,543
THAT THE PROCESS FOR THE STUDY 

1236
00:48:49,543 --> 00:48:50,778
IS WORKING PROPERLY. AGAIN THIS 

1237
00:48:50,778 --> 00:48:53,414
IS MORE ADVANCED. 

1238
00:48:53,414 --> 00:48:55,916
SO THERE'S A LOT OF FACTORS TO 

1239
00:48:55,916 --> 00:48:58,119
CONSIDER IN DESIGN SELECTION. 

1240
00:48:58,119 --> 00:48:58,719
THE NUMBER OF TREATMENTS THAT 

1241
00:48:58,719 --> 00:49:04,992
YOU HAVE. THE STUDY POOL. WHAT 

1242
00:49:04,992 --> 00:49:06,294
PEOPLE WILL TOLERATE AND THE 

1243
00:49:06,294 --> 00:49:08,763
TIMEFRAME DURATION OF THE STUDY 

1244
00:49:08,763 --> 00:49:12,400
THAT YOU CAN AFFORD TO CONDUCT. 

1245
00:49:12,400 --> 00:49:17,138
CARRYOVER EFFECTS LIKE I SAID. 

1246
00:49:17,138 --> 00:49:17,772
DRUGS WILL TAKE SLOW ACROSS THE 

1247
00:49:17,772 --> 00:49:19,106
DESIGN FOR EXAMPLE.

1248
00:49:19,106 --> 00:49:19,674
PATIENT EXPERIENCE OR ETHICAL 

1249
00:49:19,674 --> 00:49:22,877
CONCERNS.  SOME OF THE DESIGNS 

1250
00:49:22,877 --> 00:49:28,749
ARE GOING TO BE MORE THAN 

1251
00:49:28,749 --> 00:49:31,419
OTHERS, OR CERTAIN PATIENT 

1252
00:49:31,419 --> 00:49:32,053
GROUPS THAN OTHERS.

1253
00:49:32,053 --> 00:49:32,653
THE VARIABILITY OF THE SUBJECT 

1254
00:49:32,653 --> 00:49:34,088
POOL IS ALSO GOING TO BE A 

1255
00:49:34,088 --> 00:49:34,655
COMMENT CONSIDERATION FOR 

1256
00:49:34,655 --> 00:49:44,732
DESIGN HERE.  

1257
00:49:44,732 --> 00:49:50,638
AND FOR THE LAST PIECE OF THE 

1258
00:49:50,638 --> 00:49:51,238
LECTURE, I WANT TO TALK ABOUT 

1259
00:49:51,238 --> 00:49:51,839
STATISTICAL ANALYSES BROADLY 

1260
00:49:51,839 --> 00:49:52,273
AND THE ROLE OF THE 

1261
00:49:52,273 --> 00:49:56,644
STATISTICIAN IN THIS PROCESS.  

1262
00:49:56,644 --> 00:49:57,278
SO YOU CAN GET A SENSE FROM THE 

1263
00:49:57,278 --> 00:49:58,546
PREVIOUS PART OF THE LECTURE 

1264
00:49:58,546 --> 00:49:59,146
THAT THERE'S A LOT OF THINGS 

1265
00:49:59,146 --> 00:49:59,780
THAT GO INTO THE ANALYSIS THAT 

1266
00:49:59,780 --> 00:50:00,414
WE ARE NOT REALLY GOING TO GET 

1267
00:50:00,414 --> 00:50:04,552
INTO DETAIL HERE.  BUT THAT YOU 

1268
00:50:04,552 --> 00:50:05,186
WOULD NEED TO LEAN ON SOMEBODY 

1269
00:50:05,186 --> 00:50:07,421
TO HELP WITH DECISIONS ABOUT 

1270
00:50:07,421 --> 00:50:09,990
EVEN THINGS THAT WE DIDN'T TALK 

1271
00:50:09,990 --> 00:50:14,662
ABOUT IN DESIGN. SUPERIORITY, 

1272
00:50:14,662 --> 00:50:16,030
EQUIVALENT OR NONINFERIORITY 

1273
00:50:16,030 --> 00:50:16,664
QUESTIONS COMPARED TO THE USUAL 

1274
00:50:16,664 --> 00:50:18,632
SUPERIORITY QUESTIONS. 

1275
00:50:18,632 --> 00:50:20,234
MATCH DATA CONSIDERATIONS FOR 

1276
00:50:20,234 --> 00:50:26,006
THAT. SAMPLE SIZE CALCULATION.  

1277
00:50:26,006 --> 00:50:26,607
RANDOMIZED CALCULATIONS WHICH 

1278
00:50:26,607 --> 00:50:27,174
YOU UNDERSTAND A LITTLE BIT 

1279
00:50:27,174 --> 00:50:30,010
BETTER BY NOW HOPEFULLY.  AND 

1280
00:50:30,010 --> 00:50:30,611
STATISTICAL ANALYSIS PLAN AND 

1281
00:50:30,611 --> 00:50:34,882
HOW YOU PRESENT THE GOALS. 

1282
00:50:34,882 --> 00:50:39,553
AND I THINK IT CAN BE KIND OF A 

1283
00:50:39,553 --> 00:50:39,987
MISCONCEPTION, THAT 

1284
00:50:39,987 --> 00:50:40,621
STATISTICIANS ARE ONLY INVOLVED 

1285
00:50:40,621 --> 00:50:44,725
KIND OF FOR PROTOCOL WRITING 

1286
00:50:44,725 --> 00:50:45,359
POWER CALCULATIONS AND HELPING 

1287
00:50:45,359 --> 00:50:47,194
WITH THE ANALYSIS OF DATA, AND 

1288
00:50:47,194 --> 00:50:47,762
DISSEMINATION. HELPING YOU 

1289
00:50:47,762 --> 00:50:50,598
WRITE YOUR PAPER AT THE END.

1290
00:50:50,598 --> 00:50:54,335
BUT REALLY THIS IS PART OF A 

1291
00:50:54,335 --> 00:50:54,902
MUCH FULLER PICTURE OF GOAL 

1292
00:50:54,902 --> 00:51:04,011
RESEARCH MAP. THIS IS A NICE 

1293
00:51:04,011 --> 00:51:06,247
FIGURE THAT I USE TO TRY TO 

1294
00:51:06,247 --> 00:51:06,881
SHOW THIS A LITTLE BETTER. YOU 

1295
00:51:06,881 --> 00:51:07,348
CAN IMAGINE THE STUDY 

1296
00:51:07,348 --> 00:51:08,215
CONCEPTION ALTHOUGH IT TO 

1297
00:51:08,215 --> 00:51:12,653
ARCHIVING OF DATA. ETHICS 

1298
00:51:12,653 --> 00:51:14,221
REVIEWS. RECRUITMENT AND 

1299
00:51:14,221 --> 00:51:15,089
CONSENTING. AN ITERATIVE 

1300
00:51:15,089 --> 00:51:17,758
PROCESSES HERE. 

1301
00:51:17,758 --> 00:51:18,259
I WANTED TO TAKE A COUPLE 

1302
00:51:18,259 --> 00:51:21,996
MINUTES TO HIGHLIGHT ALL THE 

1303
00:51:21,996 --> 00:51:22,563
WAYS THAT STATISTICIANS CAN 

1304
00:51:22,563 --> 00:51:23,197
HELP DURING THE COURSE OF STUDY.

1305
00:51:23,197 --> 00:51:26,967
 AND, BE A KEY PART OF YOUR 

1306
00:51:26,967 --> 00:51:29,537
STUDY TEAM. 

1307
00:51:29,537 --> 00:51:32,440
FIRST WOULD BE STUDY DESIGN.  

1308
00:51:32,440 --> 00:51:33,073
THEY CAN HELP WITH A LOT OF THE 

1309
00:51:33,073 --> 00:51:34,308
THINGS I JUST MENTIONED. CHOICE 

1310
00:51:34,308 --> 00:51:35,743
OF OUTCOME MEASURES BASED ON 

1311
00:51:35,743 --> 00:51:39,447
YOUR STUDY GOALS.  CHOICE OF 

1312
00:51:39,447 --> 00:51:41,849
CONTROL GROUP, TIMING OF 

1313
00:51:41,849 --> 00:51:42,450
MEASUREMENTS, WHETHER WE WANT 

1314
00:51:42,450 --> 00:51:45,119
REPEATED MEASURES OR THE TIMING 

1315
00:51:45,119 --> 00:51:46,387
THAT YOU WANT YOUR VISIT TO BE. 

1316
00:51:46,387 --> 00:51:46,987
SOMETIMES CAN BE HELPFUL WITH 

1317
00:51:46,987 --> 00:51:51,091
THOSE THINGS. 

1318
00:51:51,091 --> 00:51:51,625
WHAT YOUR PRIOR DATA LOOKS 

1319
00:51:51,625 --> 00:51:52,193
LIKE. DO YOU HAVE ANY PILOT 

1320
00:51:52,193 --> 00:51:55,362
DATA? ANY DATA IN THE 

1321
00:51:55,362 --> 00:51:55,996
LITERATURE THAT CAN HELP YOU TO 

1322
00:51:55,996 --> 00:52:00,968
DESIGN OR POWER YOU STUDY? AND

1323
00:52:00,968 --> 00:52:01,735
THE CHOICE OF DESIGN WHETHER WE 

1324
00:52:01,735 --> 00:52:05,840
CAN DO A CROSSOVER DESIGN. 

1325
00:52:05,840 --> 00:52:06,474
WHETHER WE WANTED TO RANDOMNESS 

1326
00:52:06,474 --> 00:52:10,144
OR JUST A COHORT STUDY, ETC.  

1327
00:52:10,144 --> 00:52:10,711
WE CAN HELP WITH ALL THOSE 

1328
00:52:10,711 --> 00:52:12,813
SORTS OF QUESTIONS AS WELL.

1329
00:52:12,813 --> 00:52:16,150
STATISTICIANS CAN SERVE AS A 

1330
00:52:16,150 --> 00:52:18,719
PEER REVIEW PROCESS. WE CAN 

1331
00:52:18,719 --> 00:52:22,456
HELP VET IDEAS AND PROVIDE 

1332
00:52:22,456 --> 00:52:22,957
FEEDBACK AND TALK ABOUT 

1333
00:52:22,957 --> 00:52:24,158
DISABILITY. 

1334
00:52:24,158 --> 00:52:26,160
IN THIS DIAGRAM, THERE IS A 

1335
00:52:26,160 --> 00:52:29,930
PROTOCOL SECTION HERE.  THIS 

1336
00:52:29,930 --> 00:52:30,498
OCCURS BEFORE PROTOCOL BUT 

1337
00:52:30,498 --> 00:52:31,031
STATISTICIANS ALSO REVIEW 

1338
00:52:31,031 --> 00:52:33,267
PROTOCOLS IN FULL AS PART OF 

1339
00:52:33,267 --> 00:52:40,174
THE SCIENTIFIC REVIEW AT NINDS. 

1340
00:52:40,174 --> 00:52:41,509
FOR PREPARING STUDY 

1341
00:52:41,509 --> 00:52:43,611
OPTIMIZATION, WE CAN EVEN HELP 

1342
00:52:43,611 --> 00:52:44,211
WITH CONSIDERATIONS FOR DATA 

1343
00:52:44,211 --> 00:52:47,515
MANAGEMENT.  WE SEE DATA ALL 

1344
00:52:47,515 --> 00:52:48,983
THE TIME. OFTEN THERE'S 

1345
00:52:48,983 --> 00:52:49,583
MANAGERS ON YOUR TEAM TO HELP 

1346
00:52:49,583 --> 00:52:52,987
YOU.  BUT WE CAN ALSO PROVIDE 

1347
00:52:52,987 --> 00:52:55,055
INSIGHTS AS FAR AS DATA 

1348
00:52:55,055 --> 00:52:55,623
STORAGE, AND FORMATTING AS 

1349
00:52:55,623 --> 00:52:56,624
WELL.  THAT CAN BE VERY 

1350
00:52:56,624 --> 00:53:01,262
HELPFUL. 

1351
00:53:01,262 --> 00:53:01,729
IN THE PROTOCOL WRITING 

1352
00:53:01,729 --> 00:53:03,564
PROCESS, WE CAN HELP DO THAT 

1353
00:53:03,564 --> 00:53:05,266
WRITING, DEVELOP STATISTICAL 

1354
00:53:05,266 --> 00:53:07,668
ANALYSIS PLAN THAT IS DETAILED 

1355
00:53:07,668 --> 00:53:08,769
ENOUGH FOR GOOD, RIGOROUS 

1356
00:53:08,769 --> 00:53:12,806
STUDY.  WE CAN HELP CLARIFY OR 

1357
00:53:12,806 --> 00:53:13,674
WRITE OUTCOME MEASURES AND 

1358
00:53:13,674 --> 00:53:17,878
ENDPOINTS. AND OBVIOUSLY DO THE 

1359
00:53:17,878 --> 00:53:18,512
SAMPLE SIZE AND CALCULATION AS 

1360
00:53:18,512 --> 00:53:21,515
PART OF THE PROTOCOL WRITING. 

1361
00:53:21,515 --> 00:53:26,253
AND YOU KNOW, CONSENTING 

1362
00:53:26,253 --> 00:53:26,887
OBVIOUSLY, WE DO NOT DO AS MUCH 

1363
00:53:26,887 --> 00:53:30,291
OF THAT BUT THERE IS THIS BIG 

1364
00:53:30,291 --> 00:53:31,759
ITERATIVE PIECE IN THE MIDDLE 

1365
00:53:31,759 --> 00:53:32,426
THAT HAPPENS CONCURRENTLY. MOST 

1366
00:53:32,426 --> 00:53:33,460
OF THE TIME IN RECRUITMENT. 

1367
00:53:33,460 --> 00:53:36,196
WHICH IS PROTOCOL AMENDMENTS, 

1368
00:53:36,196 --> 00:53:38,933
MANAGEMENT AND MONITORING, AND 

1369
00:53:38,933 --> 00:53:39,533
THIS HAPPENS ALL THE TIME AS 

1370
00:53:39,533 --> 00:53:40,134
YOUR STUDY IS ONGOING AND WE 

1371
00:53:40,134 --> 00:53:44,038
CAN HELP WITH A LOT OF THAT 

1372
00:53:44,038 --> 00:53:46,240
TOO. AND MAYBE THERE INTERIM 

1373
00:53:46,240 --> 00:53:46,840
THINGS, OR DATA SUMMARIES OR 

1374
00:53:46,840 --> 00:53:55,849
DATA QUALITY CHECKING, 

1375
00:53:55,849 --> 00:53:56,450
ESPECIALLY IF BLINDED, AND WE 

1376
00:53:56,450 --> 00:53:57,685
CAN HELP IN ANY WAY WITHOUT. WE 

1377
00:53:57,685 --> 00:53:58,252
CAN HELP WITH ALL OF THESE 

1378
00:53:58,252 --> 00:54:00,688
THINGS AS WELL. 

1379
00:54:00,688 --> 00:54:09,029
AND OBVIOUSLY, CONDUCT THE 

1380
00:54:09,029 --> 00:54:09,530
ANALYSIS ON LOCKED DATA 

1381
00:54:09,530 --> 00:54:10,130
ACCORDING TO THE STATISTICAL 

1382
00:54:10,130 --> 00:54:11,632
ANALYSIS PLAN DEVELOPED. 

1383
00:54:11,632 --> 00:54:12,199
PRODUCING FINAL REPORTS AND 

1384
00:54:12,199 --> 00:54:19,907
ALSO SPECIFICALLY REPORTING TO 

1385
00:54:19,907 --> 00:54:20,507
CT.GOV, IS ALSO SOMETHING WE 

1386
00:54:20,507 --> 00:54:21,075
CAN HELP WITH WHICH YOU ARE 

1387
00:54:21,075 --> 00:54:22,509
MANDATED TO DO. 

1388
00:54:22,509 --> 00:54:23,544
AND WE CAN ASSIST WITH 

1389
00:54:23,544 --> 00:54:25,846
RESPONSES TO REVIEWERS, IN THE 

1390
00:54:25,846 --> 00:54:26,547
PUBLICATION PROCESS, AND 

1391
00:54:26,547 --> 00:54:29,650
IDEALLY AS A COLLABORATOR.  BUT 

1392
00:54:29,650 --> 00:54:31,452
IF YOU EVER COME TO A 

1393
00:54:31,452 --> 00:54:32,453
STATISTICAL QUESTIONS ABOUT 

1394
00:54:32,453 --> 00:54:34,355
COMMENTS YOU RECEIVE ON YOUR 

1395
00:54:34,355 --> 00:54:34,989
STUDY IN THE REVIEW PROCESS WE 

1396
00:54:34,989 --> 00:54:35,623
ARE ALSO ABLE TO HELP WITH ALL 

1397
00:54:35,623 --> 00:54:38,025
THAT.  AND EVEN WITH ARCHIVING. 

1398
00:54:38,025 --> 00:54:39,860
 WE CAN HELP DEVELOP THAT FINAL 

1399
00:54:39,860 --> 00:54:42,563
CLEAN DATA SET, INCLUDING THE 

1400
00:54:42,563 --> 00:54:45,299
FINAL DATA CLEANING AND 

1401
00:54:45,299 --> 00:54:48,969
ACCOMPANYING DATA DICTIONARIES. 

1402
00:54:48,969 --> 00:54:49,603
 ALL OF THAT HELPS TO MAKE THIS 

1403
00:54:49,603 --> 00:54:57,645
A CLEAN, CLOSED PROCESS.

1404
00:54:57,645 --> 00:54:58,212
AND I WOULD SAY, HAND-IN-HAND 

1405
00:54:58,212 --> 00:54:59,446
DATA SHARING TO JOURNALS WHICH 

1406
00:54:59,446 --> 00:55:04,084
IS BECOMING MORE COMMON.

1407
00:55:04,084 --> 00:55:05,252
THIS IS A COMMENT FROM A 

1408
00:55:05,252 --> 00:55:05,886
STATISTICIAN. YOU PROBABLY HAVE 

1409
00:55:05,886 --> 00:55:11,492
HEARD THIS BEFORE. A QUOTE FROM 

1410
00:55:11,492 --> 00:55:12,059
R.A. FISHER CONSIDERED THE 

1411
00:55:12,059 --> 00:55:15,029
FATHER OF MOTHER STATISTICS. 

1412
00:55:15,029 --> 00:55:17,598
"TO CALL IN A STATISTICIAN 

1413
00:55:17,598 --> 00:55:18,198
AFTER THE EXPERIMENT IS DONE 

1414
00:55:18,198 --> 00:55:19,099
MAY BE NO MORE THAN ASKING HIM 

1415
00:55:19,099 --> 00:55:22,169
TO PERFORM A POSTMORTEM 

1416
00:55:22,169 --> 00:55:22,803
EXAMINATION-- HE MAY BE ABLE TO 

1417
00:55:22,803 --> 00:55:23,404
SAY WHAT THE EXPERIMENT DIED 

1418
00:55:23,404 --> 00:55:26,874
OF." 

1419
00:55:26,874 --> 00:55:27,474
MAYBE IT IS A LITTLE BIT MORBID 

1420
00:55:27,474 --> 00:55:28,108
BUT THE TRUTH IS WE ARE ONLY AS 

1421
00:55:28,108 --> 00:55:28,709
HELPFUL AS WE ARE ALLOWED TO 

1422
00:55:28,709 --> 00:55:31,145
BE. AS THE COLLABORATORS.  WE 

1423
00:55:31,145 --> 00:55:31,745
REALLY DO NEED TO BE INVOLVED 

1424
00:55:31,745 --> 00:55:35,049
IN AS MUCH OF THE STUDY 

1425
00:55:35,049 --> 00:55:35,649
DEVELOPMENT IS GOOD DESIGN OR 

1426
00:55:35,649 --> 00:55:39,653
WRITING AS POSSIBLE.  MAKING 

1427
00:55:39,653 --> 00:55:40,287
SURE WE CAN HELP YOU WITH THAT 

1428
00:55:40,287 --> 00:55:40,954
AS POSSIBLE, AND MAKE YOUR 

1429
00:55:40,954 --> 00:55:44,458
TRIAL SUCCESSFUL. 

1430
00:55:44,458 --> 00:55:46,627
SO IN SUMMARY, THERE'S A LOT OF 

1431
00:55:46,627 --> 00:55:47,928
DECISIONS TO BE MADE IN TRIAL 

1432
00:55:47,928 --> 00:55:53,267
DESIGN.  THERE ARE MANY PIECES 

1433
00:55:53,267 --> 00:55:53,901
OF INFORMATION THAT YOU NEED TO 

1434
00:55:53,901 --> 00:55:54,501
PROPERLY PLAN AND DESIGN IN A 

1435
00:55:54,501 --> 00:55:55,669
TRIAL. 

1436
00:55:55,669 --> 00:55:56,236
WHAT PHASE ARE WE LOOKING AT 

1437
00:55:56,236 --> 00:55:57,404
ANSWER QUESTIONS? DO WE NEED A 

1438
00:55:57,404 --> 00:56:01,909
TRIAL AT ALL? IT IS ETHICALLY 

1439
00:56:01,909 --> 00:56:02,443
OKAY FOR OUR PATIENTS TO 

1440
00:56:02,443 --> 00:56:05,946
CONDUCT THE TRIAL? ARE WE 

1441
00:56:05,946 --> 00:56:06,447
TAKING ALL THE MEASURES 

1442
00:56:06,447 --> 00:56:08,082
POSSIBLE TO REDUCE BIAS AND 

1443
00:56:08,082 --> 00:56:10,784
IMPROVE RIGOR AND 

1444
00:56:10,784 --> 00:56:11,285
GENERALIZABILITY WHILE 

1445
00:56:11,285 --> 00:56:13,120
PROTECTING THE PATIENT?

1446
00:56:13,120 --> 00:56:13,687
AND AGAIN ALL THESE DECISIONS 

1447
00:56:13,687 --> 00:56:15,589
OF CONTROL GROUP. DO WE NEED 

1448
00:56:15,589 --> 00:56:17,991
MORE THAN ONE GROUP HERE? 

1449
00:56:17,991 --> 00:56:18,525
ENDPOINT DEFINITION. THE 

1450
00:56:18,525 --> 00:56:20,561
PRIMARY MEASURE OF INTEREST IN 

1451
00:56:20,561 --> 00:56:24,698
HOW WE ARE GOING TO DEFINE IT. 

1452
00:56:24,698 --> 00:56:25,299
HOW MANY GROUPS ARE GOING TO BE 

1453
00:56:25,299 --> 00:56:27,234
BLINDED OR MASKED IN THIS 

1454
00:56:27,234 --> 00:56:27,868
INVESTIGATION? AND HOW CAN WE 

1455
00:56:27,868 --> 00:56:30,971
DO THIS IN THE BEST WAY AS 

1456
00:56:30,971 --> 00:56:31,605
POSSIBLE AND CONSIDERATIONS ARE 

1457
00:56:31,605 --> 00:56:33,273
RANDOMIZATION AND HOW WE CAN 

1458
00:56:33,273 --> 00:56:33,874
CONDUCT THAT AND MAKE SURE WE 

1459
00:56:33,874 --> 00:56:36,410
ARE BALANCING OUR MAIN GROUPS 

1460
00:56:36,410 --> 00:56:39,713
IN A RANDOM WAY. 

1461
00:56:39,713 --> 00:56:40,247
AND CREATING A STUDY DESIGN 

1462
00:56:40,247 --> 00:56:43,417
THAT IS SUFFICIENT TO ANSWER 

1463
00:56:43,417 --> 00:56:44,051
THE STUDY QUESTIONS WHILE BEING 

1464
00:56:44,051 --> 00:56:44,651
EFFICIENT AND APPROPRIATE FOR 

1465
00:56:44,651 --> 00:56:45,152
THE DISEASE AND SUBJECT 

1466
00:56:45,152 --> 00:56:47,087
POPULATION. 

1467
00:56:47,087 --> 00:56:49,723
AND IN ALL OF THIS, TO WRAP UP, 

1468
00:56:49,723 --> 00:56:53,060
THE STUDY TEAM IS CRUCIAL. THE 

1469
00:56:53,060 --> 00:56:53,660
STUDY TEAM GOES FURTHER THAN 

1470
00:56:53,660 --> 00:56:58,198
JUST CLINICAL INVESTIGATORS. 

1471
00:56:58,198 --> 00:56:58,766
AND A LOT OF WHO WE ARE AS 

1472
00:56:58,766 --> 00:57:02,936
STATISTICIANS IS PART OF THAT 

1473
00:57:02,936 --> 00:57:03,470
STUDY TEAM INCLUDING DATA 

1474
00:57:03,470 --> 00:57:04,071
MANAGERS, QUALITY ASSURANCE, 

1475
00:57:04,071 --> 00:57:07,241
REGULATORY, PHARMACY, PARTICLE 

1476
00:57:07,241 --> 00:57:10,744
WRITING, MEDICAL RESOURCES, 

1477
00:57:10,744 --> 00:57:14,047
SUPPORTING DSMB AND OTHER 

1478
00:57:14,047 --> 00:57:14,681
SCIENTIFIC REVIEW PROCESSES. IT 

1479
00:57:14,681 --> 00:57:15,716
IS A HUGE TEAM TO HELP MAKE 

1480
00:57:15,716 --> 00:57:17,484
THESE DECISIONS TO MAKE SURE WE 

1481
00:57:17,484 --> 00:57:18,118
HAVE THE BEST OUTCOME POSSIBLE 

1482
00:57:18,118 --> 00:57:25,225
FOR YOUR TRIAL. 

1483
00:57:25,225 --> 00:57:27,294
THERE'S A COUPLE MINUTES LEFT.  

1484
00:57:27,294 --> 00:57:28,495
I CAN OPEN UP FOR QUESTIONS 

1485
00:57:28,495 --> 00:57:29,596
PEOPLE WANT TO PUT IN THE CHAT 

1486
00:57:29,596 --> 00:57:31,231
OR UNMUTE. I AM HAPPY TO 

1487
00:57:31,231 --> 00:57:39,706
ANSWER. 

1488
00:57:39,706 --> 00:57:40,174
THE RECORDINGS WILL BE 

1489
00:57:40,174 --> 00:57:43,177
AVAILABLE. AND THE SLIDES WILL 

1490
00:57:43,177 --> 00:57:43,811
LIKELY BE DISSEMINATED AT SOME 

1491
00:57:43,811 --> 00:57:45,312
POINT. YOU CAN KEEP AN EYE OUT 

1492
00:57:45,312 --> 00:57:55,489
FOR THE. 

1493
00:58:03,897 --> 00:58:04,398
I DON'T SEE ANY QUESTIONS 

1494
00:58:04,398 --> 00:58:06,800
COMING THROUGH.  THANK YOU ALL 

1495
00:58:06,800 --> 00:58:11,905
FOR JOINING TODAY.  PLEASE COME 

1496
00:58:11,905 --> 00:58:13,640
BACK AND JOIN US NEXT WEEK FOR 

1497
00:58:13,640 --> 00:58:21,215
THE CONTINUATION OF OUR 

1498
00:58:21,215 --> 00:58:21,782
STATISTICAL LECTURE SERIES 

1499
00:58:21,782 --> 00:58:22,382
HERE. PLEASE REACH OUT IF YOU 

1500
00:58:22,382 --> 00:58:32,259
HAVE ANY QUESTIONS.

1501
00:58:32,259 --> 00:58:35,529
HOW CAN WE GET HELP FOR POWER 

1502
00:58:35,529 --> 00:58:38,899
ANALYSIS?

1503
00:58:38,899 --> 00:58:39,366
YEAH. WHICH ARE TO YOUR 

1504
00:58:39,366 --> 00:58:40,767
STATISTICIAN, IF YOU NEED THAT 

1505
00:58:40,767 --> 00:58:51,278
INFORMATION LET ME KNOW. OUR 

1506
00:58:51,812 --> 00:58:52,346
NINDS STATISTICIANS-- I DON'T 

1507
00:58:52,346 --> 00:58:53,280
KNOW WHAT GROUP YOUR PART OF 

1508
00:58:53,280 --> 00:58:54,681
BUT IF YOU ARE HERE AT NINDS, 

1509
00:58:54,681 --> 00:59:05,058
WE ARE GLAD TO HELP. 

1510
00:59:10,731 --> 00:59:12,499
THANK YOU ALL. 

1511
00:59:12,499 --> 00:59:22,576
>> [END]