1
00:00:09,242 --> 00:00:10,377
So, let's talk

2
00:00:10,377 --> 00:00:13,446
a little
bit about actually managing the data.

3
00:00:13,446 --> 00:00:17,283
It really becomes a system
that is a continuous system.

4
00:00:17,283 --> 00:00:20,754
You know, the data's collected
according to the protocol.

5
00:00:20,754 --> 00:00:24,958
Hopefully, there's some internal quality
audit or checkpoints at the site.

6
00:00:24,958 --> 00:00:28,028
The data is submitted
according to the protocol.

7
00:00:28,028 --> 00:00:31,898
It's reviewed at the data center
where queries are often

8
00:00:31,898 --> 00:00:33,433
generated, can be generated.

9
00:00:33,433 --> 00:00:35,068
We'll talk about that.

10
00:00:35,068 --> 00:00:37,537
And then data's corrected and resubmitted.

11
00:00:37,537 --> 00:00:42,509
And this is kind of just this ongoing
cycle of data submission.

12
00:00:42,509 --> 00:00:46,646
And I think this says,
the picture says a lot,

13
00:00:46,646 --> 00:00:51,584
but just going to give an example
for a phase III trial.

14
00:00:51,584 --> 00:00:56,156
There are probably close to 7,500
to 9,000 pieces of data

15
00:00:56,156 --> 00:00:59,459
that might be collected
for a single patient.

16
00:01:00,393 --> 00:01:05,799
And so, if you take that times a study
that might have a sample size

17
00:01:05,799 --> 00:01:10,503
of 500 patients, that's going to be over
4 million pieces of data.

18
00:01:10,503 --> 00:01:15,909
So, kind of gives you a perspective
of the amount of work and the importance

19
00:01:15,909 --> 00:01:20,413
of collecting accurate data
when it comes to that level of effort.

20
00:01:20,413 --> 00:01:25,285
So again, I want to talk a little bit
about the investigator responsibilities

21
00:01:25,285 --> 00:01:28,555
as far as relating
to completion of the CRF.

22
00:01:28,555 --> 00:01:31,057
The good clinical practice,
or GCP, guidelines

23
00:01:31,891 --> 00:01:34,994
really specifically state
that the investigator should ensure

24
00:01:34,994 --> 00:01:38,431
the accuracy, completeness, legibility,
and timeliness of the data.

25
00:01:38,431 --> 00:01:41,901
And that would include
that all sections are completed,

26
00:01:41,901 --> 00:01:46,906
any alterations have been properly made,
and all adverse events are fully recorded,

27
00:01:46,906 --> 00:01:50,376
and they are reported
accordingly to regulatory bodies.

28
00:01:50,376 --> 00:01:52,545
So, timeliness is very important.

29
00:01:52,812 --> 00:01:57,684
I deal -- I feel,
and I try to really encourage my staff,

30
00:01:57,684 --> 00:02:02,288
to complete the data forms
as closely after the visit as possible.

31
00:02:03,423 --> 00:02:06,860
This is where
you pick up on maybe something was missed.

32
00:02:06,860 --> 00:02:07,794
Maybe you know

33
00:02:07,794 --> 00:02:11,564
it took place, and information was stated,
but it wasn't documented completely.

34
00:02:11,564 --> 00:02:15,034
You're going to remember,
they're going to remember to complete that

35
00:02:15,034 --> 00:02:20,039
and know the information, or record it,
closer to the visit than if a staff person

36
00:02:20,039 --> 00:02:24,777
is waiting to go back in three months
later and fill out that data form.

37
00:02:24,777 --> 00:02:27,981
It's a little too late to capture the
information that might have been missed.

38
00:02:29,983 --> 00:02:31,151
It also ensures

39
00:02:31,151 --> 00:02:34,254
that the information can be retrieved
or followed-up

40
00:02:34,254 --> 00:02:38,158
on while that visit's
still fresh in their mind.

41
00:02:38,158 --> 00:02:42,428
There are problems that are encountered
when CRFs are completed

42
00:02:42,428 --> 00:02:44,364
and this is very common.

43
00:02:44,364 --> 00:02:49,002
Sometimes they lack -- there might be
a lack of source documentation.

44
00:02:49,002 --> 00:02:53,273
I think one thing
that we continually found in our site

45
00:02:53,273 --> 00:02:56,776
were physicians
failing to report in the source document

46
00:02:56,776 --> 00:02:59,512
what the patient's
actual performance status was.

47
00:02:59,512 --> 00:03:02,815
Simple, but often
our auditors required that number

48
00:03:02,815 --> 00:03:07,320
instead of just a description
of the activity of the patient.

49
00:03:07,320 --> 00:03:09,389
It was a frustrating event.

50
00:03:09,389 --> 00:03:12,292
So we developed actually a source document

51
00:03:12,292 --> 00:03:16,396
that captured that,
so we could have the information.

52
00:03:16,396 --> 00:03:21,334
There can be errors in the protocol,
adhering to the protocol,

53
00:03:21,334 --> 00:03:25,872
there can be missing data, transcription
errors, lag in data entry,

54
00:03:25,872 --> 00:03:30,810
poor patient recall after adverse events,
and also just patient poor compliance.

55
00:03:30,810 --> 00:03:33,079
Queries are a day-to-day activity.

56
00:03:33,079 --> 00:03:36,216
If you've been involved
in clinical research

57
00:03:36,216 --> 00:03:39,819
much at all, particularly
as a research associate

58
00:03:39,819 --> 00:03:44,324
or a CRA, a research nurse,
they basically review the,

59
00:03:44,324 --> 00:03:49,729
you know, they -- reviewing the activity
within the data that's collected.

60
00:03:49,729 --> 00:03:54,701
It helps to clean the data,
and it helps to assess

61
00:03:54,701 --> 00:03:58,304
and resolve some inconsistent data,
missing or range

62
00:03:58,304 --> 00:04:00,540
discrepancies, deviations
from the protocol.

63
00:04:01,374 --> 00:04:06,045
And very commonly sponsors will generate
what we call data clarification

64
00:04:06,045 --> 00:04:10,717
forms, or DCFs, and they're sent
to the investigator for resolution.

65
00:04:10,717 --> 00:04:14,988
Again, this is more probably common
with the paper forms

66
00:04:14,988 --> 00:04:18,825
than the electronic because,
as I said, electronic forms

67
00:04:18,825 --> 00:04:23,496
-- the opportunity to collect
or to clarify and correct issues

68
00:04:23,496 --> 00:04:29,435
is usually provided at the time of data
entry, but not all the time.

69
00:04:31,104 --> 00:04:34,941
I really hope that any research sites do

70
00:04:34,941 --> 00:04:40,680
have as some sort of an internal process
to review their research

71
00:04:40,680 --> 00:04:44,517
before they have an external auditor
come in.

72
00:04:44,517 --> 00:04:49,289
To wait until that external auditor
comes in to find

73
00:04:49,289 --> 00:04:52,659
your problems is not a good idea.

74
00:04:52,659 --> 00:04:56,963
To be proactive, self-identify
your errors, develop corrective actions

75
00:04:56,963 --> 00:05:00,800
plans, correct your SOPs,
or update your SOPs,

76
00:05:02,101 --> 00:05:06,205
and then -- this is just really key
to preventing non-compliance.

77
00:05:06,205 --> 00:05:10,677
And it helps to then train your staff
more appropriately, as well,

78
00:05:10,677 --> 00:05:15,515
so hopefully those issues are resolved
at the time and aren't ongoing.

79
00:05:15,515 --> 00:05:19,252
Just a little bit on data
safety monitoring boards

80
00:05:19,252 --> 00:05:22,488
because I see in your agenda
all your topics

81
00:05:22,755 --> 00:05:27,093
for discussion that there is a discussion
on data safety monitoring boards.

82
00:05:27,093 --> 00:05:30,296
But I think it does play an important role

83
00:05:30,296 --> 00:05:32,432
when you're talking about data management.

84
00:05:32,432 --> 00:05:36,703
The data safety monitoring boards
are basically boards that are made up

85
00:05:36,703 --> 00:05:40,239
of experts that, for specific
usually disease categories, that review

86
00:05:40,239 --> 00:05:44,510
trial data, and look at accrual rates,
they look at adverse events

87
00:05:44,510 --> 00:05:48,781
that have been occurring,
and they also look at the data reports

88
00:05:48,981 --> 00:05:52,618
as far as
what is the interim analysis showing.

89
00:05:52,618 --> 00:05:57,023
They also then by doing
that might capture issues for trials

90
00:05:57,023 --> 00:06:01,461
if the accrual rate really slow,
they're really going to encourage

91
00:06:01,461 --> 00:06:06,299
that investigator to figure out why,
develop methods, or look at what

92
00:06:06,299 --> 00:06:10,703
the problems might be,
perhaps an amendment needs to be made

93
00:06:10,703 --> 00:06:15,141
to the study to change something
in eligibility that's prohibiting accrual.

94
00:06:15,975 --> 00:06:17,944
They look at adverse events.

95
00:06:17,944 --> 00:06:24,183
If one arm, it's a two-arm study,
and one arm is more toxic than the other,

96
00:06:24,183 --> 00:06:27,687
the data monitoring board might suggest,
depending on how

97
00:06:27,687 --> 00:06:30,823
severe those differences
might be, of shutting down

98
00:06:30,823 --> 00:06:34,727
one of the arms,
or even shutting down the study.

99
00:06:34,727 --> 00:06:37,864
And doing the interim analysis
also helps to

100
00:06:37,864 --> 00:06:40,967
assess is one arm
possibly giving better results.

101
00:06:40,967 --> 00:06:46,072
Are patients -- and they usually
write this out in their data analysis

102
00:06:46,072 --> 00:06:51,144
plan of what they will accept
as a level of change between arms.

103
00:06:51,144 --> 00:06:56,849
If one arm has reached that threshold,
they may actually shut a study arm down

104
00:06:56,849 --> 00:07:01,387
and continue maybe the other arm
or shut the study down itself.

105
00:07:01,387 --> 00:07:03,289
So, the data safety monitoring

106
00:07:03,289 --> 00:07:07,093
boards really do
play a key role in assessing trials.