E19 OPTIMISATION OF SAFETY DATA COLLECTION

Administrative draft guidance, not enacted legislation. This captured publisher edition preserves the listed document and attachments; later developments are not independently consolidated.

E19 OPTIMISATION OF SAFETY DATA COLLECTION
This draft guidance, when finalized, will represent the current thinking of the Food and Drug
Administration (FDA or Agency) on this topic. It does not establish any rights for any person and is
not binding on FDA or the public. You can use an alternative approach if it satisfies the requirements
of the applicable statutes and regulations. To discuss an alternative approach, contact the FDA staff
responsible for this guidance as listed on the title page. The draft guidance has been left in the original
International Council for Harmonisation format. The final guidance will be reformatted and edited to
conform with FDA’s good guidance practice regulation and style.
For questions regarding this draft document, contact
(CDER) Ellis Unger, M.D., 301-796-2270.

INTERNATIONAL COUNCIL FOR HARMONISATION OF TECHNICAL
REQUIREMENTS FOR PHARMACEUTICALS FOR HUMAN USE
ICH HARMONISED GUIDELINE
Optimisation of Safety Data Collection
E19
Draft version
Endorsed on 3 April 2019
Currently under public consultation
At Step 2 of the ICH Process, a consensus draft text or guideline, agreed by the appropriate
ICH Expert Working Group, is transmitted by the ICH Assembly to the regulatory
authorities of the ICH regions for internal and external consultation, according to national
or regional procedures.

2
E19
Document History
Code History Date
E19 Endorsement by the Members of the ICH
Assembly under Step 2 and release for public
consultation (document dated 28 February 2019).
3 April 2019
Legal notice: This document is protected by copyright and may, with the exception of the ICH logo, be
used, reproduced, incorporated into other works, adapted, modified, translated or distributed under a
public license provided that ICH's copyright in the document is acknowledged at all times. In case of
any adaption, modification or translation of the document, reasonable steps must be taken to clearly
label, demarcate or otherwise identify that changes were made to or based on the original document.
Any impression that the adaption, modification or translation of the original document is endorsed or
sponsored by the ICH must be avoided.
The document is provided "as is" without warranty of any kind. In no event shall the ICH or the authors
of the original document be liable for any claim, damages or other liability arising from the use of the
document.
The above-mentioned permissions do not apply to content supplied by third parties. Therefore, for
documents where the copyright vests in a third party, permission for reproduction must be obtained
from this copyright holder.

1
ICH HARMONISED GUIDELINE	2
Optimisation of Safety Data Collection	3
E19	4
5
TABLE OF CONTENTS	6
1 INTRODUCTION ............................................................................................................ 4	7
1.1 Objective of the Guideline ...................................................................................................... 4	8
1.2 Background ............................................................................................................................. 4	9
1.3 Scope of the Guideline ............................................................................................................ 4	10
2 GENERAL PRINCIPLES ............................................................................................... 5	11
2.1 Types of Data for Which Selective Safety Data Collection May be Appropriate .................. 5	12
2.1.1 Types of Safety Data Where it May be Appropriate to Limit or Stop Collection ............ 5	13
2.1.2 Types of Safety Data That Should Generally be Collected under All Circumstances .... 5	14
2.1.3 Baseline	15
Data……………………………………………………………………………….5	16
2.2 When May Selective Safety Data Collection Be Considered? ............................................... 6	17
2.2.1 Benefit-Risk Considerations for Selective Safety Data Collection.................................. 6	18
2.2.2 Extent of	19
Exposure………………………………………………………………………….7	20
2.3 Examples Where Selective Safety Data Collection May be Considered ................................ 7	21
2.4 Ensuring Patient Safety within Studies ................................................................................... 7	22
2.5 Changes in Approach to Safety Data Collection..................................................................... 8	23
2.6 Early Consultation with Regulatory Authorities ..................................................................... 8	24
3 METHODS OF IMPLEMENTATION .......................................................................... 8	25
3.1 Selective Safety Data Collection for All Patients in the Study ............................................... 9	26
3.2 Comprehensive Safety Data Collection for a Specific Subset(s) of the Population, with	27
Selective Safety Data Collection for Other Patients ........................................................................... 9	28
3.3 Comprehensive Safety Data Collection in a Representative Subset of the Population, with	29
Selective Safety Data Collection for Other Patients ........................................................................... 9	30
3.4 Comprehensive Safety Data Collection for the Initial Portion of the Study, with Selective	31
Data Collection Thereafter .................................................................................................................. 9	32
4 RELATIONSHIP WITH OTHER GUIDELINES/REGULATIONS ....................... 10	33
34

ICH E19 Guideline
4
1 INTRODUCTION	35
1.1 Objective of the Guideline	36
This Guideline is intended to provide internationally harmonised guidance on an optimised	37
approach to safety data collection in some late-stage pre-approval or post-approval studies	38
when the safety profile of a drug is sufficiently characterised. Optimisation of safety data	39
collection using a selective approach may improve the efficiency of clinical studies while	40
reducing the burden to study participants. Adoption of an internationally harmonised approach	41
to selective safety data collection may facilitate global participation in clinical studies.	42
1.2 Background	43
Regulators and industry have a shared interest in reducing the burden to study participants	44
while facilitating the conduct of studies that could yield important new medical knowledge and	45
advance public health. Although safety monitoring of patients during clinical studies remains	46
critically important, unnecessary and burdensome data collection may serve as a disincentive	47
to participation in clinical studies, e.g., frequent and time-consuming patient visits; laboratory	48
tests; and/or physical examinations.	49
Knowledge about a medicinal product’s safety profile continually evolves as safety data	50
accumulates. Throughout the course of medicinal product development and subsequently	51
while the drug is marketed, sponsors collect extensive safety-related data, including all vital	52
signs, laboratory data, and adverse events. In the later stages of drug development, and if the	53
safety profile is well-understood and documented, comprehensive collection of all safety data	54
may provide only limited additional knowledge of clinical importance. In such circumstances,	55
a more selective approach to safety data collection may be adequate and optimal, as long as the	56
study objectives and the welfare of study participants are not compromised.	57
Importantly, sponsors and investigators should ensure that routine patient care is not	58
compromised by the selective safety data collection approach outlined in this Guideline. It is	59
recognised that safety monitoring serves to protect individual study participants and will	60
continue to be performed as per standard of care.	61
1.3 Scope of the Guideline	62
This guidance is intended to apply to collection of safety data during the late-stage development	63
of medicinal products in interventional and non-interventional studies, in the post-approval	64
setting and, for specific cases, in the pre-approval setting.	65
In the pre-approval setting, comprehensive safety data collection is expected in order to	66
elucidate frequency, severity, seriousness, and dose-response of adverse events, including	67
potential differences across subsets, e.g., demographic; concomitant illnesses; and/or	68
concomitant therapy. However, even before approval of a new medicinal product, if there is	69
agreement with regulatory authorities that sufficient safety data are available or are being	70
collected in ongoing late-stage studies, selective safety data collection may be appropriate in	71
certain studies.	72

ICH E19 Guideline
5
Selective safety data collection following the principles of this Guideline does not alter	73
local/regional safety reporting requirements.	74
2 GENERAL PRINCIPLES	75
2.1 Types of Data for Which Selective Safety Data Collection May be Appropriate	76
2.1.1 Types of Safety Data Where It May be Appropriate to Limit or Stop Collection	77
1. Non-serious adverse events	78
2. Routine laboratory tests	79
3. Information on concomitant medications	80
4. Physical examinations (including vital signs)	81
5. Electrocardiograms	82
2.1.2 Types of Safety Data That Should Generally be Collected under All Circumstances	83
For the following types of events/data, comprehensive details should generally be provided to	84
allow adequate assessment of the event/data, e.g., history; associated adverse events; relevant	85
laboratory values; concomitant medications; vital signs; and/or follow-up outcome.	86
1. Deaths	87
2. Serious adverse events	88
3. Significant adverse events that led to an intervention, including withdrawal or dose	89
reduction of investigational medicinal product or addition of concomitant therapy	90
4. Marked laboratory abnormalities (other than those meeting the definition of serious)	91
5. Overdose	92
6. Pregnancies	93
7. Adverse events of special interest (if defined). These adverse events may warrant	94
collection of additional information across the entire study population to better	95
characterise these events (e.g., particular laboratory parameters; vital signs; risk	96
factors; concomitant therapies; and/or concomitant illnesses). For example, if	97
gastrointestinal haemorrhage was an adverse event of special interest, one might	98
want to proactively collect concomitant antithrombotic therapy across the entire	99
study population	100
8. Laboratory data, vital signs, electrocardiograms of special interest (if defined)	101
2.1.3 Baseline Data	102
Use of a selective safety data collection approach does not change considerations for baseline	103
data collection. Baseline data are needed to ensure that subjects meet inclusion and exclusion	104
criteria for study enrolment and are important in the assessment of safety. For example,	105
particular serious adverse events may occur more frequently in subgroups defined on the basis	106
of demographics, baseline disease characteristics, coexisting illnesses, or concomitant	107
therapies; analyses of such information can be important in considering the benefit-risk profile	108
of the drug.	109

ICH E19 Guideline
6
2.2 When May Selective Safety Data Collection Be Considered?	110
When sponsors choose to implement selective safety data collection for a clinical study, a	111
scientific justification should be provided. Factors that contribute to a determination that	112
selective safety data collection would be appropriate include:	113
1. The medicinal product has received marketing authorisation from a regulatory	114
authority for the indication under investigation	115
2. Availability of post-approval safety data and findings	116
3. The dose, dosing regimen, dosage form, route of administration and treatment	117
duration used in the previously conducted studies are comparable to the planned use	118
of the drug in the proposed study	119
4. The patient population from previously conducted studies is representative of	120
subjects in the planned study regarding demographic characteristics, underlying	121
medical conditions, concomitant drugs, and other important factors (e.g.,	122
Cytochrome P450 enzymes (CYP) metabolizer status)	123
5. Exposure in previously conducted (or ongoing, if applicable) studies that contribute	124
to the overall safety database, i.e., number exposure to drug, treatment duration	125
6. Consistency of the safety profile across previous studies	126
7. Characteristics of previous studies, e.g., study design; study conduct; adequacy of	127
safety monitoring/safety data collection; availability of protocols; statistical analysis	128
plan; and/or access to data	129
8. Knowledge of the mechanism of action of the medicinal product under study	130
9. Knowledge of the safety profile of approved drugs in the same pharmacologic class	131
The above factors should be considered in determining whether the safety of the medicinal	132
product has been sufficiently characterised to provide justification for selective safety data	133
collection in the proposed study.	134
In the pre-approval setting, selective safety data collection may be justifiable if sufficient safety	135
data are available from completed studies. Moreover, when sufficient safety data will be	136
forthcoming from one or more ongoing late-stage study(ies), selective safety data collection	137
may be appropriate for a concurrently conducted study-initiated pre-approval.	138
2.2.1 Benefit-Risk Considerations for Selective Safety Data Collection	139
It should be recognised that the contribution of non-serious adverse events to the benefit-risk	140
profile of a drug may differ depending on the indication of use and patient characteristics (e.g.,	141
age and/or cardiovascular risk factors). These factors should be considered when accepting the	142
comparability of patient populations and the applicability of selective safety data collection.	143
For example, even when safety of a drug is sufficiently characterised in a patient population	144
with advanced disease, comprehensive safety data collection in a patient population with less	145
advanced disease may be appropriate to ensure that the benefits outweigh the risks in the less	146
severely affected population.	147

ICH E19 Guideline
7
2.2.2 Extent of Exposure	148
Selective safety data collection could be considered for studies using lower doses and/or shorter	149
durations than in previous studies. Conversely, selective safety data collection would generally	150
not be acceptable if higher doses and/or longer treatment durations than previously studied are	151
planned. Nonetheless, even when exposure is greater in the planned study, there may be	152
circumstances where selective safety data collection is still appropriate, e.g., a study designed	153
to characterise infrequent serious adverse events (e.g., renal toxicity; myocardial infarction;	154
and/or stroke) associated with longer term use of the medicinal product within the labelled	155
indication; a planned five-year study when a one-year study has been completed.	156
2.3 Examples Where Selective Safety Data Collection May be Considered	157
Selective safety data collection may be appropriate in studies used to evaluate some of the	158
following objectives. These are not the only circumstances where selective safety data	159
collection may be appropriate.	160
1. New indications of approved drugs	161
2. To study additional endpoints, e.g., patient-reported outcome for symptomatic	162
improvement; quality of life; and/or outcome studies (e.g., mortality; morbidity; and/or	163
specific safety issues)	164
3. To study comparative effectiveness/efficacy	165
4. Demonstration of superiority when non-inferiority has been demonstrated	166
5. Characterisation of adverse events of special interest	167
6. Fulfilment of post-approval requirements, post-authorisation safety studies based on	168
data collection from registries or electronic health records	169
7. Late-stage premarketing outcome study in a large population	170
Additional examples and situations for applying selective safety data collection may be found	171
in Section 3, Methods of Implementation.	172
2.4 Ensuring Patient Safety within Studies	173
Patient safety monitoring serves two purposes: 1) to protect the welfare of individual study	174
participants; and 2) to accumulate safety information to be used in the assessment of benefit-	175
risk for the proposed indication. The recommendations in this Guideline do not obviate the	176
need for monitoring to protect individual patient welfare. Although certain safety data, e.g.,	177
non-serious adverse events, would not need to be recorded in the case report form (CRF) when	178
selective safety data collection is determined to be appropriate, the protocol should stipulate	179
that patients are monitored per standard of care. For example, for a medicinal product known	180
to cause hyperglycaemia, where routine blood glucose monitoring is recommended in labeling,	181
glucose should be monitored in patients participating in a study. If hyperglycaemia is well-	182
characterised with this medicinal product, the glucose data do not need to be recorded in the	183
CRF or reported to the sponsor in studies using selective safety data collection. Glucose levels	184
would be recorded in the CRF and reported to the sponsor if stipulated in the protocol, e.g., as	185
an adverse event of special interest, associated with a serious adverse event.	186

ICH E19 Guideline
8
2.5 Changes in Approach to Safety Data Collection	187
When an unexpected safety issue arises during the course of a study, e.g., a postmarketing	188
safety signal; a finding from a nonclinical study; higher than expected withdrawals; and/or	189
concern from a data monitoring committee; a change in the selective safety data collection	190
approach may be warranted, e.g., denoting a new adverse event of special interest; and/or	191
reverting to comprehensive safety data collection.	192
2.6 Early Consultation with Regulatory Authorities	193
Studies must be conducted according to local and regional laws and regulatory requirements.	194
When sponsors are considering selective safety data collection in interventional studies, they	195
should discuss their scientific rationale and planned methods with regulatory authorities prior	196
to initiating the study(ies). The same applies to non-interventional studies that are being	197
conducted to address requests from regulatory authorities.	198
It is possible to conduct a multi-regional clinical study using a single protocol with selective	199
safety data collection if the safety profile of the product is considered to be sufficiently	200
characterised, and all regulatory authorities agree with the proposed approach. A well-	201
designed multi-regional clinical study that takes this Guideline into account will help the	202
sponsor reach agreement with regulatory authorities in multiple regions (See ICH E17 –	203
General Principles for Planning and Design of Multi-Regional Clinical Trials).	204
3 METHODS OF IMPLEMENTATION	205
Having considered the principles outlined in Section 2, General Principles, with respect to	206
when it may be appropriate to limit or stop collection of certain types of safety data, a number	207
of approaches for selective safety data collection may be considered.	208
Use of selective safety data collection can introduce important complexities in study conduct	209
and safety analysis. The specific approaches should be carefully planned and clearly delineated	210
within the relevant study documents, e.g., protocol; monitoring plan; and/or statistical analysis	211
plan, with a reference to this Guideline.	212
Regardless of the method chosen, it is essential to ensure patient safety and adhere to local and	213
regional laws and regulations. When the selective safety data collection approach is used for	214
a clinical study, the approach should be described in the appropriate document(s) when safety	215
findings are presented, e.g., the Clinical Study Report (CSR); Development Safety Update	216
Report (DSUR); Periodic Benefit-Risk Evaluation Report (PBRER); Periodic Safety Update	217
Report (PSUR); and/or Common Technical Document (CTD).	218
The following examples of methods of implementation are not meant to be all-inclusive. These	219
approaches can be applied in both the pre- and post-approval settings and require a scientific	220
rationale and justification. The data supporting these approaches are more likely to be available	221
in the post-approval setting than in the pre-approval setting.	222

ICH E19 Guideline
9
3.1 Selective Safety Data Collection for All Patients in the Study	223
For all patients in the study, parameters listed in Section 2.1.2, General Principles, are collected	224
throughout the study, e.g., serious adverse events; adverse events of special interest; and/or	225
deaths. Conversely, the parameters listed in Section 2.1.1, General Principles, are not	226
collected, e.g., non-serious adverse events; routine laboratory values; concomitant	227
medications; physical examination data; vital signs; and/or electrocardiograms.	228
In the post-approval setting, this approach may be useful to address a specific safety concern,	229
for example, to meet a post-authorisation commitment, when safety in other regards has been	230
sufficiently characterised.	231
In the pre-approval setting, this approach may be also used. For example, consider a	232
development programme for a lipid-lowering drug, where a decrease in low-density lipoprotein	233
(LDL) cholesterol will serve as the basis of approval, but the impact on cardiovascular risk is	234
being investigated. In addition to the completed Phase 2 programme, two Phase 3 studies are	235
ongoing with LDL cholesterol as the primary endpoint, which will provide adequate exposure	236
to assess safety sufficiently. The sponsor wishes to initiate a third study with major adverse	237
cardiovascular events as the primary endpoint. For the third study, a selective safety data	238
collection approach could be justified considering the data available in light of the principles	239
above.	240
3.2 Comprehensive Safety Data Collection for a Specific Subset(s) of the Population,	241
with Selective Safety Data Collection for Other Patients	242
Comprehensive safety data are collected for specific subset(s) of the patient population where	243
additional information is deemed important, whereas selective safety data are collected for	244
other patients. For example, if the patient population in previous studies included few patients	245
over the age of 65, it could be of value to collect full data on this population in a new study in	246
the same indication or in a related indication. Other examples of specific subsets include those	247
based on geographic location; ethnicity; sex; baseline disease status (renal/hepatic impairment),	248
CYP status; or genetics.	249
3.3 Comprehensive Safety Data Collection in a Representative Subset of the	250
Population, with Selective Safety Data Collection for Other Patients	251
In some cases, efficacy studies must enrol many thousands of patients in order to achieve	252
adequate statistical power. In such settings, such as a large clinical outcomes study, the number	253
of patients planned for enrolment may greatly exceed the number needed to assess the non-	254
serious adverse events adequately. In this setting, comprehensive safety data could be collected	255
for only a representative subset of patients, for example, full data collection could be	256
undertaken at randomly selected sites.	257
3.4 Comprehensive Safety Data Collection for the Initial Portion of the Study, with	258
Selective Data Collection Thereafter	259
Comprehensive safety data are collected from baseline through some pre-determined interval	260
of the study, with selective safety data collection thereafter. A data monitoring committee	261

ICH E19 Guideline
10
could consider the safety data and provide agreement with selective safety data collection for	262
the subsequent portion of the study. These approaches can be useful for studies designed to	263
assess important long-term drug effects, where safety would be adequately characterised in the	264
early part of the study, e.g., one year, through comprehensive safety data collection. For	265
example, consider a study to prevent an important outcome such as dementia, end-stage kidney	266
disease, and/or hepatic failure. Assuming it would take three years to collect adequate events	267
to have adequate statistical power for efficacy, it may be appropriate to utilize a selective	268
approach to safety data collection once data have been analysed for all patients followed	269
through one year and non-serious adverse events have been deemed to be adequately	270
characterised. The selective approach would discontinue collection of non-serious adverse	271
events, vital signs, laboratory tests, etc., and utilize less frequent study visit intervals. The	272
protocol should include a prospective plan for concurrence of a data monitoring committee	273
prior to the change to selective safety data collection.	274
4 RELATIONSHIP WITH OTHER GUIDELINES/REGULATIONS	275
This guideline should be considered in conjunction with other ICH guidelines relevant to the	276
conduct of clinical studies and clinical safety data management, e.g., E2A (Clinical Safety Data	277
Management: Definitions and Standards for Expedited Reporting); E2F (Development Safety	278
Update Report); E3 (Structure and Content of Clinical Study Reports ); E6(R2) (Good Clinical	279
Practice: Integrated Addendum to ICH E6(R1)); E8 (General Considerations for Clinical	280
Trials); and/or E17 (General Principles for Planning and Design of Multi-Regional Clinical	281
Trials). Evaluation of the information generated through post-approval pharmacovigilance	282
activities is also important for all products to ensure their safe use, e.g. E2E (Pharmacovigilance	283
Planning); E2D (Post-Approval Safety Data Management: Definitions and Standards for	284
Expedited Reporting); and E2C(R2) (Periodic Benefit-Risk Evaluation Report).	285