Document text
COVID -19 m RNA Vaccine
C4591009 NON -INTERVENTIONAL STUDY PROTOCOL SYNOPSIS
Version 1.0, 01 April 2021
PFIZER CONFIDENTIAL
Page 1of 25Study information
Title A Non -Interventional Post- Approval Safet y
Study of the Pfizer -BioNTech COVID -19
mRNA V accine in the United States.
Protocol number C4591009
Protocol synopsis version identifier 1.0
Date 01 April 2021
EU Post Authoriz ation Study, PAS
register numberTo be registered before the start of data
collection
Active substance COVID -19 mRNA Vaccine is single -
stranded, 5 '-capped messenger RNA
(mRNA )produced using a cell -free in vitro
transcription from the corresponding DNA
templates, encoding the viral spike (S)
protein of SARS- CoV -2.
Medicinal product Pfizer -BioNTech COVID- 19 Vaccine
Research question and objectives Research question: What are the incidence
rates /prevalence of safet y events of interest
among individuals vaccinated with the
Pfizer -BioNTech COVID-19 vaccine within
selected United States data sources
participating in the Sentinel Sy stem, overall
and in subpopulations of interest (pregnant
women , immu nocompromised individuals,
and individuals with a history of COVID -19)
compared with rates of those events in
individuals who have not received an y
vaccination for COVID -19?
090177e196b004bd\Approved\Approved On: 13-Apr-2021 20:49 (GMT)
FDA-CBER-2021-5683-0780098
COVID -19 m RNA Vaccine
C4591009 NON -INTERVENTIONAL STUDY PROTOCOL SYNOPSIS
Version 1.0, 01 April 2021
PFIZER CONFIDENTIAL
Page 2of 25Primary objectives
To estimate the relative risk (RR) or
prevalence ratio of safety events of
interest following receipt of at least one
dose of the Pfizer -BioNTech COVID -19
vaccine within the overall study
population
To estimate the RRor prevalence ratio of
safet y events of interest following receipt
of at least one dose of the Pfizer -
BioNTech COVID -19 vaccine in
pregnant women, in
immunocompromised individuals, and in
individuals with a history of COVID -19
Secondary objectives
To describe the proportion of individuals
receiving at least one dose and a
complete dose series of the Pfizer -
BioNTech COVID -19 vaccine, within
the overall stud y population, in pregnant
women , in immunocompromised
individuals, and in individuals with a
history of COVID -19
To describe —among indivi duals who
receive a first dose of the Pfizer -
BioNTech COVID -19 vaccine— the
timing and type of second dose of
COVID -19 vaccine ( Pfizer -BioNTech
COVID -19 vaccine or other COVID -19
vaccine ), within the overall study
population , in pregnant women, in
immunocompromised individuals, and in
individuals with a history of COVID -19
To describe baseline characteristics
(demographics and comorbidities )of
individuals who receive at least one dose
of the Pfizer -BioNTech COVID-19
vaccine an d those who receive no
COVID -19 vaccine of any type, within
the overall stud ypopulation , in pregnant
women, in immunocompromised
090177e196b004bd\Approved\Approved On: 13-Apr-2021 20:49 (GMT)
FDA-CBER-2021-5683-0780099
COVID -19 m RNA Vaccine
C4591009 NON -INTERVENTIONAL STUDY PROTOCOL SYNOPSIS
Version 1.0, 01 April 2021
PFIZER CONFIDENTIAL
Page 3of 25individuals, and in individuals with a
history of COVID -19
Author Alison Kawai , ScD
RTI Health Solutions
Waltham, Massachusetts
Jeffrey Brown, PhD
Department of Population Medicine
Harvard Medical School & Harvard Pilgrim
Health Care Institute
Boston, Massachusetts
Cynthia de Luise, MPH, PhD
Risk Management and Safet y Surveillance
Research
Pfizer, I nc.
New York , New York
090177e196b004bd\Approved\Approved On: 13-Apr-2021 20:49 (GMT)
FDA-CBER-2021-5683-0780100
COVID -19 m RNA Vaccine
C4591009 NON -INTERVENTIONAL STUDY PROTOCOL SYNOPSIS
Version 1.0, 01 April 2021
PFIZER CONFIDENTIAL
Page 4of 25TABLE OF CONTENTS
APPENDI CES ................................ ................................ ................................ ........................... 5
1. TI TLE ................................ ................................ ................................ ................................ ....6
2. RATIONALE AND BAC KGROUND ................................ ................................ .................. 6
3. RESEARCH QUESTION AND OBJECTI VES ................................ ................................ ...6
4. STUDY DESIGN ................................ ................................ ................................ ................... 7
5. POPUL ATION ................................ ................................ ................................ ...................... 7
6. VARIABLES ................................ ................................ ................................ ......................... 8
6.1. Safet y events ................................ ................................ ................................ ............. 8
6.2. Vaccine exposures ................................ ................................ ................................ ...10
6.3. Covar iates ................................ ................................ ................................ ................ 10
7. DATA SOURCES ................................ ................................ ................................ ............... 11
7.1. CVS Health, Aetna ................................ ................................ ................................ ..13
7.2. HealthCore ................................ ................................ ................................ .............. 13
7.3. HealthPartners ................................ ................................ ................................ ......... 13
7.4. Humana ................................ ................................ ................................ ................... 13
7.5. Optum Research Database ................................ ................................ ...................... 13
8. STUDY SIZE ................................ ................................ ................................ ....................... 14
9. DATA ANALYSI S................................ ................................ ................................ .............. 15
9.1. Descriptive anal ysis................................ ................................ ................................ .15
9.2. Com parative anal ysis................................ ................................ .............................. 16
9.2.1. Overall approach ................................ ................................ ......................... 16
9.2.2. Analy ses in overall study population, immunocompromised
individuals, and individuals with history of COVID -19................................ ..16
9.2.3. Analy ses in the pregnant population................................ ........................... 17
9.3. Further details on comparative analy sis................................ ................................ ..18
9.4. Sensitivity analy ses................................ ................................ ................................ .18
9.4.1. Self -controlled risk interval design................................ ............................. 18
9.4.2. Cohort design with historical comparators................................ ................. 19
9.4.3. Alternative risk intervals ................................ ................................ ............. 19
9.5. Monitoring and interim analy sis................................ ................................ ............. 19
10. STRENGTHS AND LI MITATIONS ................................ ................................ ................ 20
090177e196b004bd\Approved\Approved On: 13-Apr-2021 20:49 (GMT)
FDA-CBER-2021-5683-0780101
COVID -19 m RNA Vaccine
C4591009 NON -INTERVENTIONAL STUDY PROTOCOL SYNOPSIS
Version 1.0, 01 April 2021
PFIZER CONFIDENTIAL
Page 5of 2511. MILESTONES ................................ ................................ ................................ ................... 21
12. REFERENCES ................................ ................................ ................................ .................. 23
APPENDICES
Appendix 1. Investigator list ................................ ................................ ................................ ....24
090177e196b004bd\Approved\Approved On: 13-Apr-2021 20:49 (GMT)
FDA-CBER-2021-5683-0780102
COVID -19 m RNA Vaccine
C4591009 NON -INTERVENTIONAL STUDY PROTOCOL SYNOPSIS
Version 1.0, 01 April 2021
PFIZER CONFIDENTIAL
Page 6of 251.TITLE
A Non -Interventional Post-A pproval Safet y Study of the Pfizer -BioNTech COVID-19
mRNA V accine in the United States
2.RATIONALE AND BACKGROUND
The novel coronavirus SARS -CoV -2, the cause of COVID -19, has resulted in a global
pandemic. On December 11, 2020 ,the Pfizer -BioNTech COVID-19 vaccine was authorized
for emergency use by the Food and Drug Administration (FDA ) to prevent COVID- 19 in
individuals aged 16years and older in the United States (US) (FDA, 2020 ). Pfizer -BioNTech
is submitting a Biologic al License Application ( BLA ) for marketing approval of the vaccine
for the prevention of SARS -CoV -2 infection in individuals age d12 years and older .
Post-authorization observational studies using real- world data are needed to assess the
association between the Pfizer -BioNTech COVID-19 vaccine and pre- determined safety
events of interest inindividuals administered the vaccine in the general population and in
subpopulations of interest (e.g., pregnant women, immunocompromised individuals, and
individuals with a history of COVID -19). This protocol sy nopsis describes a proposed
observational study of safety events of interest occurring in recipients of the Pfizer -
BioNTech COVID -19 vaccine using claims data and electronic health records data (where
available ) from Data Partners participating in the Sentinel Sy stem. The safety events of
interest in this study are based on those included in COVID -19 vaccine safety surveillance in
the FDA 's Biologics Effectiveness and Safety (BEST )System (Wong et al., 2021 )and the
Centers for Disease Control (C DC)Vaccine Safety Datalink (VSD ) (Shimabukuro et al.,
2021 ). Additional safet y events of interest may be added as new evidence develops during
the pandemic .This protocol sy nopsis summarizes a proposed non- interventional study ,
designated as a post -authorization safety study ,which is anticipated as a commitment to the
FDA.
3.RESEARCH QUESTION AND OBJECTIVES
Research question: What are the incidence rates /prevalence of safet y events of interest
among individuals vaccinated with the Pfizer -BioNTech COVID-19 vaccine within selected
USdata sources participating in the Sentinel Sy stem, overall and in subpopulations of
interest (pregnant women , immunocompromised individuals, and individuals with a history
of COVID -19), compared with rates of those events in individuals who have not received an y
vaccination for COVID -19?
Primary objectives
To estimate the relative risk(RR) or prevalence ratio of safet y events of interest
following receipt of at least one dose of the Pfizer -BioNTech COVID -19 vaccine within
the overall stud y population
To estimate th e RRor prevalence ratio of safet y events of interest following receipt of at
least one dose of the Pfizer -BioNTech COVID -19 vaccine in pregnant women, in
immunocompromised individuals, and in individuals with a history of COVID -19
090177e196b004bd\Approved\Approved On: 13-Apr-2021 20:49 (GMT)
FDA-CBER-2021-5683-0780103
COVID -19 m RNA Vaccine
C4591009 NON -INTERVENTIONAL STUDY PROTOCOL SYNOPSIS
Version 1.0, 01 April 2021
PFIZER CONFIDENTIAL
Page 7of 25Secondary objectives
To describe the proportion of individuals receiving at least one dose and a complete dose
series of Pfi zer-BioNTech COVID- 19 vaccine, within the overall study population ,in
pregnant women, in immunocompromised individuals, and in individuals with a history
of COVID -19
To describe —among persons who receive a first dose of the Pfizer -BioNTech COVID -19
vaccine —the timing and ty pe of second dose of COVID- 19 vaccine (Pfizer -BioNTech
COVID -19 vaccine or other COVID -19 vaccine), within the overall study population , in
pregnant women, in immunocompromised individuals, and in individuals with a history
of COVI D-19
To describe baseline characteristics (demographics and comorbidities )of individuals who
receive at least one dose of the Pfizer -BioNTech COVID- 19 vaccine and those with no
record of COVID -19 vaccination of an y type , within the overall study populatio n, in
pregnant women, in immunocompromised individuals, and in individuals with a history
of COVID -19
4.STUDY DESIGN
This isa retrospective cohort study comparing vaccinated individuals with concurrent
unexposed comparator s. Vaccinated individuals will be matched to concurrent unexposed
comparators (in a ratio of at least 1:2) on data source and calendar time for anal ysis in the
overall study population , immunocompromised individuals, and individuals with a history of
COVID -19. In pregnant women, t hose who are vaccinated will be matched to concurrent
unexposed comparators (in a ratio of at least 1:2) on maternal age and pregnancy start.
Propensity score methods will be used to adjust for confounding through matching or in
regression anal ysis. The st udy will use data from five Data Partners that participate in the
Sentinel Sy stem .
The study period will start onthe da tethat the Pfizer -BioNTech COVID -19 vaccine was
granted emergency use authorization in the US (December 11, 2020 )and will end a
minimum of 3 years after this date .
5.POPULATION
The source population for this study will be commercial health plan enrollees from five Data
Partners that contribute claims and electronic health records data to the Sentinel Sy stem :
Aetna /CVS Health, HealthC ore/Anthem , HealthPartners, Humana, and
Optum/ UnitedHealthcare .
Individuals of all ageswill be included in the descriptive anal ysis of Pfizer -BioNTech
COVID -19vaccine utilization .
Safety anal ysis is planned to be limited to individuals within the age -indicated population for
the Pfizer -BioNTech COVID-19 vaccine. However, if the number of individuals receiving
the vaccine outside the age -indicated range is substantial (criteria for determining this to be
090177e196b004bd\Approved\Approved On: 13-Apr-2021 20:49 (GMT)
FDA-CBER-2021-5683-0780104
COVID -19 m RNA Vaccine
C4591009 NON -INTERVENTIONAL STUDY PROTOCOL SYNOPSIS
Version 1.0, 01 April 2021
PFIZER CONFIDENTIAL
Page 8of 25defined in the statistical analysis plan [SAP] ), then these individuals will be included in the
safet y analysis.
Individuals will be eligible for the study if they have continuous medical and pharmacy
insurance coverage for at least 1 2months before baseline (i.e.,the start of follow -up).
Women will be eligible to be included in anal ysis of the pregnant population if they were
pregnant for at least 1 day during the study period (regardless of the timing of pregnancy
start). Anal ysis of congenital malformations, preterm birth, and small size for gestational age
will be limited to pregnancies ending in a livebirth ; linked data on infants during the first
year of life will be used to identify select safet y events of interest .Additional study eligibility
criteria are described in the data analysis section (Section 9).
6.VARIABLES
6.1.Safety events
Safety events of interest will be identified in claims and electronic health records (where
available, as not all Data Partners will have access to electronic health records) using
diagnosis codes, with procedure and/orpharmacy dispensing codes as appropriate .Detailed
definitions will be included in the SAP.
Outcomes likely to be miscla ssified (to be determined based on clinical expert opinion and
review of prior validation studies if available) may be validated. As needed, validation of
select outcomes will be performed by clinicians without knowledge of vaccination status by
review of medical records or patient profile s (i.e.,chronological listings of codes in claims
data).
The following safet y events of interest will be assessed in the general population,
immunocompromised individuals, individuals with a history of COVID -19,and pregnant
women . This list comprises events being monitored in rapid cy cle anal ysisof COVID -19
vaccines inthe FDA 's BEST System (Wong et al., 2021 )and the CDC 's VSD(Shimabukuro
et al., 2021 ), with the addition of vaccine -associated enhanced respiratory disease :
Acute disseminated encephalomy elitis
Acute m yocardial infarction
Acute respiratory distress sy ndrome
Anaph ylaxis
Appendicitis
Bell's pals y
Convulsions/seizures
Disseminated intravascular coagulation
090177e196b004bd\Approved\Approved On: 13-Apr-2021 20:49 (GMT)
FDA-CBER-2021-5683-0780105
COVID -19 m RNA Vaccine
C4591009 NON -INTERVENTIONAL STUDY PROTOCOL SYNOPSIS
Version 1.0, 01 April 2021
PFIZER CONFIDENTIAL
Page 9of 25Encephalom yelitis
Guillain -Barré syndrome
Thrombotic thrombocy topenic purpura
Immune thrombocy topenia
Kawasaki disease
Multi inflammatory syndrome
Myocarditis/pericarditis
Narcoleps y
Stroke, hemorrhagic
Stroke, ische mic
Transverse m yelitis
Deep vein thrombosis
Vaccine -associated enhanced respiratory disease
Venous thromboembolism
Pulmonary embolism (subset of venous thromboembolism )
The following pregnancy outcomes will be assessed in pregnant women or their infants:
Spontaneous abortion (spontaneous pregnancy loss before 20 completed weeks gestation)
Stillbirth (fetal deaths at or after 20 completed weeks gestation)
Preterm birth
Major congenital malformations
Small size for gestational age
Other emergent safet y events of interest may be added as the understanding of the safet y
profile of the Pfizer -BioNTech COVID-19 vaccine evolves and feasibility of their
assessment permits in the data sources .
090177e196b004bd\Approved\Approved On: 13-Apr-2021 20:49 (GMT)
FDA-CBER-2021-5683-0780106
COVID -19 m RNA Vaccine
C4591009 NON -INTERVENTIONAL STUDY PROTOCOL SYNOPSIS
Version 1.0, 01 April 2021
PFIZER CONFIDENTIAL
Page 10of 25Risk windows will be defined for safet y events of interest that have a hypothesized increased
risk during specific time periods following vaccination . For other safet y events of interest ,
patients will be followed for safety events for a maximum of 1 year.
6.2.Vaccine exposures
Exposures to the Pfizer -BioNTech COVID -19 vaccine will be identified in claims and
electronic health records data via pharmacy dispensing and/ or procedure codes. It is
anticipated that some vaccinations administered outside traditional medical care settings may
not be capt ured in claims or electronic health records data . Therefore, sensitivity anal ysis will
incorporate stud y designs that are less susceptible to misclassification of unexposed status
(Section 9.4).For anal ysis of pregnancy outcomes , except for congenital malformations,
exposures occurring an ytime during pregnancy (excluding those that occur after the at-risk
period for the outcome where applicable , e.g., in analy sis of preterm birth, a woman
vaccinat edafter 37- weeks gestation would not be considered exposed ) will be considered.
For congenital malformations, only pregnancies with vaccinations occurring during the
exposure window ( i.e.,first trimester ) will be considered exposed.
6.3.Covariates
The following potential confounders will be identified:
Demographics: age, sex, and race/ethnicit y (iffeasible ). Data on race/ethnicity are
anticipated to be incomplete, and the feasibility of including this variabl e in the anal yses
will be assessed before study start
Comorbidities, identified in claims or electronic health records data (where available, as
not all Data Partners will have access to diagnosis , procedure , and dispensing codes from
electronic health re cords) in the 1 2months before the index date : history of anaph ylaxis,
history of allergies, diabetes (type 1and t ype2), hypertension, cardiovascular disease,
cerebrovascular diseases, chronic respiratory disease, chronic kidney disease, chronic
liver disease, cancer, autoimmune disorders, influenza and other respiratory infections .
Comorbidities will be identified via diagnosis codes (with p rocedure and/or pharmacy
dispensing codes as appropriate ).
Medications and non –COVID -19 vaccinations in the 1 2 mon ths before the index date
(including vaccines administered concomitantl y with the Pfizer -BioNTech COVID -19
vaccine ), identified in claims or electronic health records data via procedure and/or
pharmacy dispensing codes
Health care utilization in the 1 2 mon ths before the index date: number of hospitalizations ;
number of emergency department visits, cancer screening (s); skilled nursing facility ,
nursing home ,or extended care facility stay; other preventive health care services, as
appropriate
Immunocomprom ising conditions (to be considered as a potential confounder in anal ysis of
the overall study population and to identify cohorts for safety and descriptive anal ysis in
090177e196b004bd\Approved\Approved On: 13-Apr-2021 20:49 (GMT)
FDA-CBER-2021-5683-0780107
COVID -19 m RNA Vaccine
C4591009 NON -INTERVENTIONAL STUDY PROTOCOL SYNOPSIS
Version 1.0, 01 April 2021
PFIZER CONFIDENTIAL
Page 11of 25immunocompromised individuals) will be identified using diagnosis, procedure, and
dispensing codes for immunodeficiencies, immunosuppressant medication use, human
immunodeficiency virus and other immunosuppressing conditions, and receipt of organ or
bone marrow transplants.
Pregnancy status (to be considered as a potential confounder in the anal ysis of the overall
study population and to identify cohorts for safet y and descriptive analysis in pregnant
women )will be identified in claims or electronic health records data.An algorithm th at
incorporates diagnosis and /orprocedure codes willidentif ythe final pregnancy outcome as
well as the start and end dates for each pregnancy episode.
History of COVID -19 before the index date (to identify cohorts for safet y and descriptive
analysis in individuals with prior history of COVID -19)will be identified in claims or
electronic health records data via diagnosis codes.
7.DATA SOURCES
This study will u sedata from five Data Partners, including data from four national insurers
(CVS Health/Aetna, HealthCore/Anthem, Humana, and Optum/UnitedHealthcare )and one
regional insurer (HealthPartners ). Each Data Partner is a participant in the FDA Sentinel
System. These data sources capture longitudinal medical care information on outpatient
medication dispe nsings, vaccine administrations, and inpatient and outpatient diagnoses and
procedures. The data sources also capture member demographic and health plan enrol lment
information. Each Data Partner canrequest access to full -text medical records for outcome
validation for at least a subset of participants. All Data Partners are able to link to external
data sources such as state immunization registries and can collect additional information via
survey s in at least a subset of members . As part of their participation in Sentinel ,three Data
Partners (CVS Health, HealthCore, and Optum) maintain a mother -infant linkage table to
support studies of medication exposures during pregnancy .As all Data Partners contribute
data to the Sentinel Sy stem, this study will leve rage Sentinel data and distributed query ing
infrastructure, including quality -checked and curated data formatted to the Sentinel Common
Data Model ( SCDM) and the publicly available Sentinel analy tic tools (Curtis et al., 2012;
Sentinel, 2018).
The Sentinel Sy stem is an active surveillance s ystem that uses routine query ing and
analytical tools to evaluate electronic health care data from a distributed data network for
monitoring the safety of regulated medical products in the US , established under the Sentinel
Initiative (Behrman et al .,2011 ; Platt et al. , 2018 ). The average enrollment length for patients
across data sources in Sentinel is similar to that in other claims databases of members with
medical and pharmacy coverage ; approximately 25% of patients have over 3 y ears of
enrollment, and patients with chronic conditions such as diabetes and older members
typicall y have longer than average enrollment periods within these databases.
The Data Partners use the SCDM ( Curtis et al., 2012 ; Sentinel, 2018) for standardization of
demographic and clinical data elements. Publicly available routine anal ytical tools (i.e.,
reusable, modular Statistical Anal ysis Sy stem [SAS] programs ) designed to be executed
090177e196b004bd\Approved\Approved On: 13-Apr-2021 20:49 (GMT)
FDA-CBER-2021-5683-0780108
COVID -19 m RNA Vaccine
C4591009 NON -INTERVENTIONAL STUDY PROTOCOL SYNOPSIS
Version 1.0, 01 April 2021
PFIZER CONFIDENTIAL
Page 12of 25against the SCDM permit rapid and standardized querie s across partners, including
descriptive analy ses and complex methodologies (e.g., comparative anal yses).
Specific information in the SCDM includes, but is not limited to, the following t ypes of data:
Enrollment data, including one record per covered indiv idual per unique enrollment span.
Individuals are assigned a unique identifier b y their insurer that is linkable to all other
data in the SCDM. Each record in the enrollment file indicates the patient identifier,
enrollment start and end dates, and whether the patient was enrolled in medical coverage,
pharmacy coverage, or both during that range.
Demographic data, including birth date, sex, race/ethnicity , and ZI P code of their most
recentl y recorded primary residence.
Outpatient pharmacy dispensing data, including the date of each vaccination or
prescription dispensing, the National Drug Code , NDC identifier associated with the
dispensed product, the nominal day s’ suppl y, and the number of individual units (e.g.,
pills, table ts, vials )dispensed. Products purchased over the counter or at some cash -only
retail locations selling prescription drug products (e.g., through the Walmart Prescription
Program )are not consistently captured.
Medical encounter data, including the health care provi der most responsible for the
encounter as well as the facility atwhich the encounter occurred and its ZIP code.
Admission and discharge dates (if applicable )are also included in addition to the
encounter t ype(i.e., an ambulatory visit, Emergency Departm ent visit, inpatient hospital
stay, non -acute inpatient stay , or otherwise unspecified ambulatory visit ). Discharge
disposition (i.e., alive, expired, or unknown )as well as discharge status (i.e., to where a
patient was discharged ) are also included for i npatient hospital stay s and non- acute
inpatient stay s.
Diagnosis data, including the date of diagnosis, its associated encounter identifier,
admission date, provider identifier, and encounter ty pe. Diagnoses are recorded with
ICD-9-CM, International Class ification of Diseases, 9th Revision, Clinical Modification
and I CD-10-CM, International Classification of Diseases, 10th Revision, Clinical
Modification codes. For inpatient hospital and non -acute inpatient stay encounters, the
SCDM includes both principal and non- principal discharge diagnosis data.
Procedure data, including the procedure date (e.g., date of vaccination) , its associated
encounter identifier, admission date, provider identifier, and encounter t ype,are coded as
ICD-9-CM and ICD-10- Procedure Coding S ystem procedure codes; Current Procedural
Terminology , CPT categories II, III, or IV codes ; revenue codes andHealthcare Common
Procedure Coding S ystem(HCPCS )levels II and III codes.
090177e196b004bd\Approved\Approved On: 13-Apr-2021 20:49 (GMT)
FDA-CBER-2021-5683-0780109
COVID -19 m RNA Vaccine
C4591009 NON -INTERVENTIONAL STUDY PROTOCOL SYNOPSIS
Version 1.0, 01 April 2021
PFIZER CONFIDENTIAL
Page 13of 25A brief description of each individual data source is below.
7.1.CVS Health , Aetna
Aetna, a CVS Health company , is one of the nation 's leading health care benefits companies,
serving 3 8million people. Aetna became an FDA Sentinel Data Partner in 2008. Aetna 's
SCDM captures longitudinal information on dispensed prescriptions, inpatient and outpatient
diagnoses, and inpatient and outpatient treatments and procedures.
7.2.HealthCore
HealthCore, Inc., a participant in Sentinel since 2008, is a significant contributor to t he
Sentinel Collaboration and Sentinel Distributed Database.
As of February 2021, there were 79 million unique individuals with medical coverage and
approximately 60million with medical and pharmacy coverage available for research.
7.3.HealthPartners
HealthPa rtners is the largest consumer -governed nonprofit health care organization in the
US, providing care, insurance coverage, research, and education to its members.
HealthPartners serves more than 1. 8million medical and dental health plan members and
more than 1.2 million patients. HealthPartners Institute has been a member of the Sentinel
Initiative since 2008.
7.4.Humana
Humana/Comprehensive Health Insights (CHI) is a health economics and outcomes research
subsidiary of Humana Pharmacy Solutions, which focuses on treatment effectiveness, drug
safet y, adherence, medical and pharmacy benefit design, disease management programs, and
other health care services based on the Humana health plan member population.
Humana/CHI has been an active collaborator and Data Part ner in the FDA Sentinel Sy stem.
Humana databases represent geographic coverage for the entire US population and represent
over 2 7million lives.
7.5. Optum Research Database
The Optum Research Database (ORD )isa proprietary research database that contains
eligibility data and medical claims and includes health plan members who are geographicall y
diverse across the USand comprise approximately 3% to 4% of the US population. Optum
has curated and qualit y-checked data formatted to the SCDM available for use and is a
longtime participant in the Sentinel Sy stem.
090177e196b004bd\Approved\Approved On: 13-Apr-2021 20:49 (GMT)
FDA-CBER-2021-5683-0780110
COVID -19 m RNA Vaccine
C4591009 NON -INTERVENTIONAL STUDY PROTOCOL SYNOPSIS
Version 1.0, 01 April 2021
PFIZER CONFIDENTIAL
Page 14of 258.STUDY SIZE
Assuming that the true RR = 1 and a matching ratio of 1:2, the table below presents the
probability that the upper limit of the 95% confidence interval (CI)for the observed RRwill
be below 1.5, 2.0, 2.5, and 3.0 for study sizes ranging from 5 00,000 to 20,000,000 vaccinated
individuals (1,000,000 doses to 40,000,000 doses, under the assumpt ion that each person will
receive 2 doses ). These estima tes are presented to cover a range of safet y events of interest
with respect to rareness, based on background rates in the general population.
Safety
event of
interestEstimat
ed
backgr
ound
rate per
100,000
person-
years
(Black
et al.,
2021 )Number
of
individu
als
vaccinat
edProbability that the upper confidence limit of RR will be below the following
thresholdsa:
1.5 2.0 2.5 3.0
Guillain -
Barré
syndrome1.68 500,000 0.07 0.12 0.18 0.24
1,000,00
00.10 0.20 0.31 0.42
2,500,00
00.18 0.42 0.64 0.80
5,000,00
00.31 0.70 0.91 0.98
10,000,0
000.54 0.94 1.00 1.00
20,000,0
000.83 1.00 1.00 1.00
Bell's palsy 25.2 500,000 0.31 0.70 0.91 0.98
1,000,00
00.54 0.94 1.00 1.00
2,500,00
00.90 1.00 1.00 1.00
5,000,00
01.00 1.00 1.00 1.00
10,000,0
001.00 1.00 1.00 1.00
20,000,0
001.00 1.00 1.00 1.00
090177e196b004bd\Approved\Approved On: 13-Apr-2021 20:49 (GMT)
FDA-CBER-2021-5683-0780111
COVID -19 m RNA Vaccine
C4591009 NON -INTERVENTIONAL STUDY PROTOCOL SYNOPSIS
Version 1.0, 01 April 2021
PFIZER CONFIDENTIAL
Page 15of 25Safety
event of
interestEstimat
ed
backgr
ound
rate per
100,000
person-
years
(Black
et al.,
2021 )Number
of
individu
als
vaccinat
edProbability that the upper confidence limit of RR will be below the following
thresholdsa:
1.5 2.0 2.5 3.0
Myocardial
infarction 208 500,000 0.99 1.00 1.00 1.00
1,000,00
01.00 1.00 1.00 1.00
2,500,00
01.00 1.00 1.00 1.00
5,000,00
01.00 1.00 1.00 1.00
10,000,0
001.00 1.00 1.00 1.00
20,000,0
001.00 1.00 1.00 1.00
RR = relative risk.
a. Estimates in this table assume a risk window duration of 42 days for Guillain -Barré syndrome, and 28 days for
Bell's palsy and myocardial infarction.
9.DATA ANALYSIS
All analy ses will be conducted separately within each data source. Pooled analysis of RR and
prevalence ratio estimates from all data sources will be conducted using meta- analy sis
techniques or other appropriate anal ytic techniques.
Detailed methodology for summary and statistical anal yses of data collected in this study will
be documented in theSAP, which will be dated, filed, and maintained b y the sponsor. The
SAP may modify the plans outlined in the protocol ; any major modifications of primary
endpoint definitions or their anal yses willbe reflected in a protocol amendment.
9.1.Descriptive analysis
Descriptive anal ysis will be conducted in order to report on utilization of the Pfizer -
BioNTech COVID -19 vaccine during the overall study period and during the study period ,
stratified in 12 -week increments. The proportion of individuals receiving at least one dose
and a complete dose series of the Pfizer- BioNTech COVID -19 vaccine will be estimated
within the overall study population, in pregnant women, in immuno compromised individuals,
and in individuals with a history of COVID -19. Among patients who receive a first dose of
the Pfizer -BioNTech COVID- 19 vaccine, the proportion of patients will be reported by type
090177e196b004bd\Approved\Approved On: 13-Apr-2021 20:49 (GMT)
FDA-CBER-2021-5683-0780112
COVID -19 m RNA Vaccine
C4591009 NON -INTERVENTIONAL STUDY PROTOCOL SYNOPSIS
Version 1.0, 01 April 2021
PFIZER CONFIDENTIAL
Page 16of 25of second dose of COVID- 19 vaccine and time between the two doses will be described
using summary measures.
9.2.Comparative analysis
9.2.1. Overall approach
Because the Pfizer -BioNTech COVID -19 vaccine is currentl y recommended in a series of
two doses given 3 weeks apart, separate exposed cohorts will be formed for e ach dose. The
dose 1 and dose 2 cohorts will be matched to concurrent unexposed comparators (ratio of at
least 1:2) on calendar time for safet y comparative analyses in the general population,
immunocompromised individuals, and individuals with a history ofCOVID -19; and on
maternal age and pregnancy start for safet y comparative anal ysis of pregnant women.
Unexposed comparators may contribute to the exposed cohorts if they subsequently receive
the Pfizer -BioNTech C OVID-19 vaccine during the study period. Separate anal yses will be
conducted for each safety event. Individuals with safet y events of interest in a pre -specified
washout period before the index date will be excluded to ensure that incident events are
identified during the study period. Confounding bias will be addressed with propensity scores
through matching or in regression analysis. Data from the dose 1 and dose 2 cohorts and their
matched comparators will be combined , if appropriate, to obtain incidence rates/proportions
and RR/incid ence proportions of safet y events following receipt of at least one dose of the
Pfizer -BioNTech COVID- 19 vaccine.
9.2.2. A nalyses in overall study population, immunocompromised individuals , and
individuals with history of COVID -19
For analy sis of safet y events inthe overall study population, individuals receiving each
vaccine dose will be matched to unexposed concurrent comparators within Data Partner on
time period -specific propensity scores within 1 -month periods of calendar time .The index
date (i.e.,start of follow -up)in the vaccinated cohorts will be the date of vaccination.
In unexposed comparators, the index date selected willbewithin close temporal proximity
(e.g., within the same calendar month )to the date of vaccination in the corresponding
exposed persons . Individuals will be eligible to be selected as unexposed comparators if they
have no record of COVID-19 vaccination before the potential index date. Estimating
propensity scores and matching within narrow time intervals accounts for the changing
predictors of being vaccinated over time, seasonality (i.e.,temporal trends) of SARS -CoV -2
and other respiratory infections, and changes in health care utilization over time. Follow -up
in both groups will begin on or the day after the index date (depending onthe safet y event of
interest )and end at the earliest of the following: end of the study period, end of data
availability , disenrollment from the health plan, death, occurrence of the safety event of
interest, end of the duration of the outcome- specific risk window (or 1 year for outcomes
without a known risk window) , or receipt of a dose of the Pfizer -BioNTech COVID-19
vaccine or an y other COVID- 19 vaccine. If dose 2 is the Pfizer -BioNTech COVID-19
vaccine, individuals will stop follow -up in the dose 1 cohort and will start follow -up in the
dose 2 cohort . Immunocompromised individuals and individuals with a history of COVID -19
before the index date will be drawn and anal yzedseparately from the general population
using the same approach.
090177e196b004bd\Approved\Approved On: 13-Apr-2021 20:49 (GMT)
FDA-CBER-2021-5683-0780113
COVID -19 m RNA Vaccine
C4591009 NON -INTERVENTIONAL STUDY PROTOCOL SYNOPSIS
Version 1.0, 01 April 2021
PFIZER CONFIDENTIAL
Page 17of 259.2.3. A nalyses in the pregnant population
An algorithmic approach will be used to identify pregnancies in women of reproductive age,
as well as pregnancy start and end dates. Each vaccine dose during pregnancy will be
matched to unexposed concurrent comparator pregnancies (matching ratio of at least 1:2)
within Data Partner s on maternal age and estimated pregnancy start date (i.e., date of last
menstrual period [LMP]).
For analy sis of all outcomes in pregnant women ,except for congenital malformations, the
index date (i.e.,start of follow -up) in the vaccinated cohorts ( dose 1 and dose 2 cohorts )will
be the date of vaccination. The index date in unexposed comparators will be assigned tothe
equivalent of the gestational age at vaccination of the corresponding exposed comparators.
For analy sis of non -pregnancy –related outcomes, pregnancies will be eligible to be
unexposed comparators if they have no record of COVID -19 vaccin ation before the potential
index date . For anal ysis of pregnancy outcomes ,except congenital malf ormations,
pregnancies will be eligible to be included in the unexposed comparator group if they have
no record of COVID -19 vaccination anytime during pregnancy before the potential index
date or within the 42 days before pregnancy start ; women who have received any COVID -19
vaccine more than 42 days before their pregnancy begins will be eligible to be included in
analyses of these outcomes. Matching on pregnancy start date adjusts for seasonalit y
(i.e.,temporal trends )of SARS -CoV -2 and ot her respiratory infections as well as changes in
health care utilization over time. Matching exposed and unexposed pregnancies on
gestational age at the start of follow -up avoids bias due to the changing underly ing risk of
pregnancy outcomes over the cours e of pregnancy.
Confounding will be adjusted for in regression analy sis with the use of propensity scores.
Follow -up for non-pregnancy –related outcomes in pregnant women and for pregnancy
outcomes (except congenital malformations, small size for gestational age, and preterm birth)
will begin on the index date or the day after the index date (depending on the safet y event of
interest )and end at the earliest of the following: end of the study period, end of data
availability , disenrollment from the health plan, death, occurrence of the safety event of
interest, end of the duration of the outcome- specific risk window (or 1 year for outcomes
without a known risk window) , receipt of the Pfizer -BioNTech COVID- 19 vaccine or any
other COVID -19 vaccine , orend of pregnancy (for anal ysis of pregnancy outcomes). If dose
2 is the Pfizer -BioNTech COVID- 19 vaccine, individuals will stop follow -up in the dose 1
cohort and will continue follow -up in the dose 2 cohort. Small size for gestational age and
preterm bir th will be identified at birth or shortly after birth in the mother or infant.
For analy sis of congenital malformations, each exposed pregnancy will be matched to
unexposed comparator pregnancies (matching ratio of at least 1:2)within Data Partner s on
maternal age and estimated pregnancy start date. The index date (i.e.,cohort entry date and
baseline period for defining covariate s)in both the exposed and unexposed cohorts will be
pregnancy start. The prevalence of congenital malformations in infants born to mothers with
exposure to at least one dose of Pfizer-BioNTech COVID- 19 vaccine during the exposure
window ( e.g., first trimester of pregnancy )will be compared with those born tomothers with
no exposure to any COVID -19 vaccine during the exposure window. Propensity scores will
090177e196b004bd\Approved\Approved On: 13-Apr-2021 20:49 (GMT)
FDA-CBER-2021-5683-0780114
COVID -19 m RNA Vaccine
C4591009 NON -INTERVENTIONAL STUDY PROTOCOL SYNOPSIS
Version 1.0, 01 April 2021
PFIZER CONFIDENTIAL
Page 18of 25be used to anal ytically adjust for confounding. The time period for identify ing congenital
malformations in the infant will start at birth and will end at age 1 year, disenrollment from
the health plan, diagnosis of the outco me, death, end of data availability , or end of the study
period, whichever is earliest.
9.3.Further details on comparative analysis
For the cohort design, the distribution of demographics, comorbidities, and other potential
confounders will be reported and com pared between the matched exposed and unexposed
cohorts. Balance in the matched cohorts will be assessed using standardized differences or
other suitable methods.
Before comparative analy sis of the matched cohorts, propensity scores will first be estimated
by Data Partner sseparately in the overall study population , pregnant women ,
immunocompromised individuals, and individuals with a history of COVID -19,as the
probability of being vaccinated vs. the probability of being in the comparator cohort at the
index date using logistic regression and baseline variables ; propensit y scores will be
calculated in distinct time periods (e.g., within the same calendar month )to account for
changes in predictors of vaccination over time.
Incidence rates (for all outcomes except pregnancy -related outcomes )or
incidence/prevalence proportions ( for pregnancy -related outcomes) and 95% CIs of each of
the safet y events of interest will be estimated separatel y for the matched exposed and
unexposed cohorts.
For comparative analyses of all outcomes except pregnancy -related outcomes, Cox models or
Poisson regression will be used to estimate hazard ratios or incidence rate ratios and 95% CIs
within the propensity score -matched cohorts . For comparative anal ysis of pregnancy
outcomes, logistic regression will be used to estimate prevalenc e or incidence proportion
ratios and 95 %CIs. Comparative anal ysis in pregnant women will analytically adjust for
propensity score. Risk/prevalence or incidence rate differences (depending on the safet y
event of interest )and 95 %CIs will be estimated for all safet y events of interest .Because
each dose of vaccine within the same person will be considered a separate observation but
combined in anal ysis, the estimation of variance will account for the correlation between
dose 1 and dose 2 using appropriate statistical methods.
9.4.Sensitivity analyses
9.4.1. Self-controlled risk interval design
Sensitivity analy ses incorporating a SCRI design will be implemented for acute outcomes
with known risk periods in the overall stud y population, pregnant women (non-pregnancy –
related outcomes onl y), immunocompromised individuals, and individuals with a history of
COVID -19. Pregnancy -related outcomes will not be anal yzed with the SCRI design. Onl y
vaccinated individuals wil l be included in the SCRI anal ysis; the rate of a specific safet y
event in a post -vaccination risk interval will be compared with the rate in a control interval
within the same person. The control interval will be the same length as the risk interval and
will comprise person -time after the dose 2 risk interval ; these intervals will be outcome
090177e196b004bd\Approved\Approved On: 13-Apr-2021 20:49 (GMT)
FDA-CBER-2021-5683-0780115
COVID -19 m RNA Vaccine
C4591009 NON -INTERVENTIONAL STUDY PROTOCOL SYNOPSIS
Version 1.0, 01 April 2021
PFIZER CONFIDENTIAL
Page 19of 25specific and specified in more detail in the SAP . A washout period between the risk and
control intervals may be incorporated for safet y events for which the risk inte rval is not well
known. Because of the self -controlled nature of the design, bias of RR estimates arising from
differences in the distribution of time -constant confounding factors between vaccinated and
unvaccinated individuals is avoided with the SCRI des ign. Furthermore, as the design only
includes vaccinated individuals, it avoids the potential for misclassification of unexposed
status due to incomplete capture of COVID -19 vaccinations in claims or electronic health
record data.
9.4.2. Cohort design with histor ical comparators
If feasible, sensitivity analy sis of non- acute safet y outcomes in the general population and in
pregnant women will incorporate historical comparator cohorts of influenza vaccinees from a
time period before the introduction of COVID -19 vac cines. The use of a historical unexposed
comparator avoids the potential for misclassification of unexposed status due to incomplete
capture of COVID -19 vaccinations in claims or electronic health record data. The feasibility
of this analy sis will depend on the absence of trends in coding for each safety event of
interest over time in the historical comparator period and the study period.
9.4.3. Alternative risk intervals
For events for which the risk intervals are not well known (to be defined in the SAP),
descriptive analy ses of the timing of events relative to vaccination will be conducted . If
temporal clusters of increased risk following vaccination are identified , sensitivity analy ses
will be conducted using alternative risk intervals.
9.5.Monitoring and inte rim analysis
Before the final safety analy sis, results from the monitoring anal ysis will be reported in
Q3 2022 . The monitoring reports will describe the number and proportion of total individuals
who have received the Pfizer -BioNTech COVID -19 vaccine in t he overall study population,
in immunocompromised individuals, and in individuals with a history of COVID -19. The
completeness of exposure data and the number of vaccinated individuals outside the age
indication will also be assessed in monitoring analy ses.
Results from an interim analy sis will be reported in Q3 2023 . The results from the interim
analysis will include those reported from the monitoring analy sis.Additionally , the
distribution of characteristics in the vaccinated and unvaccinated individuals and incidence
rates of each safet y event of interest (overall, not by exposure status) will be reported in the
overall study general population, individuals with a history of COVID -19, and
immunocompromised individuals. Notably , pregnant women will only be analy zed as a
separate population in the final anal ysis to allow sufficient time for data onpregnancies to
accrue, as pregnancies will be identified by their outcomes (e.g., livebirth, stillbirth,
spontaneous abortion) and the data lag is approximately 6to 9 months for most Data
Partners .
090177e196b004bd\Approved\Approved On: 13-Apr-2021 20:49 (GMT)
FDA-CBER-2021-5683-0780116
COVID -19 m RNA Vaccine
C4591009 NON -INTERVENTIONAL STUDY PROTOCOL SYNOPSIS
Version 1.0, 01 April 2021
PFIZER CONFIDENTIAL
Page 20of 2510.STRENGTHS AND LIMITATIONS
A major strength of this study isthat it will include a very sizeable source population in the
US, as the participating Data Partners together collect data on more than 100 million
individuals. The use of secondary data will enable the efficient assessment of several safet y
events of interest identified by the CDC’s VSD and the FDA’s BEST Initiative, in addition to
pregnancy -related safet y events of interest in pregnant women , while using robust study
design and anal ytic approaches to adjust for potential confound ing.Moreover , the secondary
use of administrative data collected as part of routine medical care will avoid recall bias ; this
approach alsoavoids selection bias that might occur in primary data collection studies, as a
patient’s inclusion in this study is not voluntary .
Nevertheless , this study is subject to limitations arising from the use of secondary data and
the selected study design s. Limitations related to the data sources include the potential for
lack of recording in claims and electronic health records of COVID -19 vaccines administered
without reimbursement from health insurer s. This situation may lead to misclassification of
expose d individuals as “unexposed” comparators , which will underestimate vaccine coverage
rates andbias comparative risk estimates for the cohort design with concurrent unexposed
comparator s. The completeness of exposure data will be assessed in monitoring analyses
before the end of the study; if the data appear to be substantially incomplete, then the primary
study design may be reconsidered (for example, the SCRI and the cohort design with
historical unexposed comparators may be designated as the primary study designs ; and/or
linkage to immunization registries may be considered). Additionall y, the use of claims data
and electronic health records data may lead to some misclassification of outcomes (e.g., false
positives and false negatives) . Some events, suc h as spontaneous abortion , will be
incompletely captured in existing databases. Conversel y, there have been limited validation
studies of I CD-10-CMbased algorithms for safety events of interest, andthe accuracy of
algorithms for man y safety events of int erest are unknown. When possible, validated
algorithms will be use d, and o utcomes that are likely to be misclassified (based on prior
validation studies and clinical expert input) may be validated through review of medical
records or claims profiles, depen ding on the safet y event of interest .
A study design- related limitation of both the cohort and SCRI designs is that any uncertainty
regarding risk periods will lead to misclassification and attenuation of risk estimates.
Sensitivity analy ses with alternative risk intervals will be considered for outcomes for which
the risk interval is not well known.
A limitation of the cohort design is the potential for residual or unmeasured confounding
because it is unlikel y that the d ata sources will have information on all potential confounders.
To address potential confounding, the SCRI , which automatically adjusts for time -invariant
confounders, will be used as a secondary approach where feasible. However, the SCRI is not
well suite d to study outcomes with gradual onset, long latency , or risk periods that are not
well known. The SCRI may also be subject to bias for outcomes that affect exposure
probability .
A limitation specific to the cohort design with concurrent unexposed compara torsis that
unvaccinated individuals may become exposed to the COVID -19 vaccine at any time during
090177e196b004bd\Approved\Approved On: 13-Apr-2021 20:49 (GMT)
FDA-CBER-2021-5683-0780117
COVID -19 m RNA Vaccine
C4591009 NON -INTERVENTIONAL STUDY PROTOCOL SYNOPSIS
Version 1.0, 01 April 2021
PFIZER CONFIDENTIAL
Page 21of 25thestudy;if this situation occurs frequentl y, the amount of unexposed person- time in the
unexposed comparator group will be substantially reduced, which will limit the precision of
RRestimates. Forming twoseparate exposed cohorts by dose number and matching
unexposed to exposed at the time of each v accine will minimize the loss of unexposed
person -time due to receipt of vaccine in these individuals between the first and second doses.
Additionally , the sensitivity anal yses with the historical unexposed comparator cohort and
the SCRI willnot be subject to this limitation.
11.MILESTONES
Below is a proposed schedule of milestones, subject to discussion wi th the FDA.
Milestone Planned date Description of milestone
Registration in the EU
PAS registerTBD To be registered before the start of data collection .
Start of data collection ,
estimatedaQ2 2022 Start of data collection is the planned date for
starting data extraction for the purposes of the
study analysis .
Monitoring Analysis 1
Reportb,cQ3 2022 Vaccine counts and proportion sof individuals in
the databases who were vaccinated, w ithin the
overall study population, in immunocompromised
individuals, and in individuals w ith a history of
COVID -19.
Interim Study Report Q3 2023 Vaccine counts and proportion sof individuals in
the databases who were vaccinated, w ithin the
overall study population, in immunocompromised
individuals, and in individuals w ith a history of
COVID -19.
Distribution of characteristics among exposed and
unexposed individuals within the overall study
population, in immunocompromised individuals,
and in individuals with a history of COVID -19.
Incidence rates of safety events of interest, overall ,
without regard to exposure status in the overall
study population, in immunocompromised
individuals, and in individuals w ith a history of
COVID -19.
End of data collection Q1 2025 End of data collection is the planned date on which
the analytical data setwill first be completely
available . The analytic aldata setis the min imum
set of data required to perform the statistical
analysis for the study objectives.
090177e196b004bd\Approved\Approved On: 13-Apr-2021 20:49 (GMT)
FDA-CBER-2021-5683-0780118
COVID -19 m RNA Vaccine
C4591009 NON -INTERVENTIONAL STUDY PROTOCOL SYNOPSIS
Version 1.0, 01 April 2021
PFIZER CONFIDENTIAL
Page 22of 25Milestone Planned date Description of milestone
Final Study Report Q3 2025d Descriptive analysis of vaccine utilization in the
overall study population, in immunocompromised
individuals, in individuals with a history of
COVID -19, and in pregnant women .
Comparative safety analysis in the overall study
population, in immunocompromised individuals, in
pregnant w omen, and in individuals with a history
of COVID -19.
BLA = Biological License Application ; PAS = post -authorization study ; TBD = to be determined.
a.If BLA approved in Q3 2021.
b.Only includes Data Partners with a data lag of < 6months.
c.Monitoring counts will not incorporate enrollment or any other study eligibility criteria.
d.Report may be delayed to Q4 202 5, depending on the extent of validation and/or the need for external
linkages.
090177e196b004bd\Approved\Approved On: 13-Apr-2021 20:49 (GMT)
FDA-CBER-2021-5683-0780119
COVID -19 m RNA Vaccine
C4591009 NON -INTERVENTIONAL STUDY PROTOCOL SYNOPSIS
Version 1.0, 01 April 2021
PFIZER CONFIDENTIAL
Page 23of 2512.REFERENCES
Behrman RE, Benner JS, Brown JS, et al. Developing the Sentinel Sy stem --a national
resource for evidence development. N Engl J Med. 2011;3646:498- 99. doi:
10.105 6/NEJMp1014427 [published Online First: 2011/01/14].
Black SB, Law B, Chen RT, Dekker CL, Sturkenboom M, Huang WT, et al. The critical role
background rates of possible adverse events in the assessment of COVID -19 vaccine safet y.
Vaccine. 2021. doi: https://doi.org/10.1016/j.vaccine.2021.03.016 .
Curtis L H, Weiner MG, Boudreau DM, O'Cooper W, Daniel G, Nair Vet al. Design
considerations, architecture, and use of the Mini- Sentinel distributed data system.
Pharmacoepidemiol Drug Saf. 2012;2 (Suppl 1) :23-31. doi:10.1002/pds.2336 [published
Online First: 2012/01/25].
Platt R, Brown JS, Robb M, McClellan M, Ball R, Ngu yen MD, et al. The FDA Sentinel
Initiative —an evolving national resource. N Engl J Med. 2018;37922:2091 -93.
Sentinel [I nternet]. Silver Spring, MD. Food and Drug Administration, FDA c2010. 2018.
Distributed Database and Common Data Model. Available at:
https://www.sentinelinitiative.org/sentinel/data/distributed- database -common -data-model .
Accessed March 11, 2021.
Shimabukuro, T. Advisory Committ ee on Immunization Practices, ACI P: COVID -19
Vaccine Safet y Update. 2021. Available at:
https://www.cdc.gov/vaccines/acip/meetings/downloads/slides -2021 -02/28 -03-01/05 -covid-
Shimabukuro.pdf . Accessed March 5, 2021.
U.S. Food and Drug Administration (FDA). FDA Takes Key Action in Fight Against
COVID -19 By Issuing Emergency Use Authorization for First COVID -19 Vaccine. 2020.
Available at: https://www.fda.gov/news -events/press -announcements/fda- takes -key-action-
fight -against -covid-19- issuing -emergency -use-authorization -first-covid- 19. Accessed March
5, 2021.
Wong HL, Zhou CK, Thompson D, Dimova R, Clarke T, Forshee R, et al. COVID -19
Vaccine Safet y Surveillance: Active Monitoring Master Protocol. 2021. Available at:
https://www.bestinitiative.org/wp -content/upl oads/2021/02/C19 -Vaccine -Safet y-Protocol -
2021.pdf . Accessed March 5, 2021.
090177e196b004bd\Approved\Approved On: 13-Apr-2021 20:49 (GMT)
FDA-CBER-2021-5683-0780120
COVID -19 m RNA Vaccine
C4591009 NON -INTERVENTIONAL STUDY PROTOCOL SYNOPSIS
Version 1.0, 01 April 2021
PFIZER CONFIDENTIAL
Page 24of 25Appendix 1. Investigator list
Principal Investigators and Contributors to the Protocol Synopsis
Nam e, degrees Job t itle Affiliation
Cynthia de Luise, MPH, PhD Senior Director, Epidemiology Pfizer
Alison Kawai, ScD Senior Research Epidemiologist RTI Health Solutions
Jeffrey Brown, PhD Associate Professor in the
Department of Population
MedicineHarvard Medical School &
Harvard Pilgrim Health Care
Institute
J Bradley Layton, PhD Senior Rese arch Epidemiologist RTI Health Solutions
Catherine Panozzo, PhD Assistant Professor in the
Department of Population
MedicineHarvard Medical School &
Harvard Pilgrim Health Care
Institute
Alicia Gilsenan, PhD, FISPE Senior Director and Head,
EpidemiologyRTI Health Solutions
Brian Calingaert, MS Director, Epidemiology Analysis RTI Health Solutions
Catherine Johannes, PhD Senior Director, Epidemiology RTI Health Solutions
Data Partner Coordinating Investigators
Nam e, degrees Job t itle Affiliation
Cheryl N McMahill -Walraven,
MSW, PhDDirector, Safety & Collaboration CVS Health
Aziza Jamal -Allial, PhD Senior Epidemiologist HealthCore
Kevin Haynes, PharmD, MSCE Principal Scientist HealthCore
Pamala A. Pawloski, P harm D.,
BCOP, FCCPSenior Research Investigator HealthPartners Institute
Vinit Nair, BPharm, MS, RPh Principal & Director ,
Government Research &
ConsortiumsHumana
Florence Wang, ScD Vice President, Epidemiology Optum
Jessica Franklin, PhD Principal Consultant,
EpidemiologyOptum
090177e196b004bd\Approved\Approved On: 13-Apr-2021 20:49 (GMT)
FDA-CBER-2021-5683-0780121
COVID -19 m RNA Vaccine
C4591009 NON -INTERVENTIONAL STUDY PROTOCOL SYNOPSIS
Version 1.0, 01 April 2021
PFIZER CONFIDENTIAL
Page 25of 25Nam e, degrees Job t itle Affiliation
Note: Data Partner Coordinating investigators have reviewed and contributed to this protocol.
090177e196b004bd\Approved\Approved On: 13-Apr-2021 20:49 (GMT)
FDA-CBER-2021-5683-0780122