Monitoring Vaccine Safety


Safety standards for vaccines are very high given they are used among healthy persons for prevention rather than treatment and it is often difficult to predict who will be exposed to and develop a particular disease, if they are not immune. Additionally, vaccines often target vulnerable populations such as infants and pregnant individuals. Vaccines can be required for day care, school entrance and attendance, and certain professions such as healthcare workers, increasing the government’s burden to monitor and ensure safety. Lastly, vaccines are often used among a very large proportion of the population, such as all children, and consequently a rare risk of adverse reactions may impact a significant number of people. There are a broad range of activities conducted by the federal government, the private sector and academic investigators to optimize the safety of vaccines and identify, characterize and minimize vaccine adverse reactions when they occur. These activities are described below, largely summarizing a more detailed review that is available online 1.


Pre-clinical Studies

Pre-clinical safety assessments are laboratory and animal studies that do not include humans. These studies characterize products by physical, chemical and biological methods, control of the manufacturing process, and lot release tests for safety, purity, and potency 2. As vaccine candidates are developed, they are examined for possible adverse events. Ingredients in vaccines such as any adjuvants, which are substances that boost the body’s immune response to vaccines, are evaluated. Nonclinical studies are carried out in animal models where potential toxicity is assessed in regard to dosage, route of administration, number of doses and other characteristics. Vaccines targeting pregnant individuals or females of reproductive age include assessment of adverse effects on fetal development in animals.

After nonclinical assessments have provided evidence the vaccine is safe to test in humans, the sponsor (academic institution, organization or manufacturer developing the vaccine) submits an Investigational New Drug (IND) application to the Food and Drug Administration (FDA). The FDA evaluates the information on vaccine toxicology, manufacturing and quality control, plans for clinical studies, and the investigators’ clinical trials expertise. After review the FDA grants permission to initiate clinical trials in humans; however, the FDA has the authority to stop the study at any time if safety concerns arise 3.

Clinical Trials

Following FDA approval, the investigator must obtain Institutional Review Board (IRB) approval to ensure that appropriate safeguards are put in place to protect all study subjects during the trial. Clinical trials are then conducted in four phases summarized below:

Phase I: 20-100 healthy subjects are included to assess dose-related toxicity and immunogenicity.

Phase II: 50-1,000 healthy subjects are divided into control and intervention groups to assess vaccine responses, safety, and sometimes clinical efficacy. Also, dose-related immune responses are usually assessed.

Phase III: 1,000-50,000 or more subjects. Similar to Phase II trial but with greater power to assess safety due to the larger study size. Most phase III trials assess actual clinical efficacy or an immune correlate of protection if known, and the proportion of vaccinees who develop that immune protection level. Phase II trials are also used to collect data on vaccine production consistency and manufacturing scale up. Data generated from Phase III trials are the major trials upon which FDA evaluates products for safety and effectiveness to determine if a license is warranted.

Phase IV: These are conducted after licensure and are more likely to be able to detect delayed onset and rarer adverse events.

Clinical trials are phased to maximize the information gained while minimizing the risks to human subjects. Phase I and II usually include only healthy volunteers and may include only a restricted age range. Phase III trials target population for the use of the vaccine post-licensure and may include individuals with underlying health problems. Each trial usually has a Data and Safety Monitoring Board (DSMB) composed of an independent panel of experts who review safety data and can call for the end of the trial due to unacceptable risks or because it is deemed unethical to not provide the vaccine to those in the control group as the data conclusively show the vaccine is both effective and safe. Clinical trials in each phase may involve a true placebo or the experimental vaccine may be compared to an existing licensed vaccine.

Regulatory Approval and Vaccine Licensure

The sponsor, usually the manufacturer, submits a Biologics License Application (BLA) to the FDA for vaccine licensure. A BLA includes biological and chemical compositions of the vaccine, results from clinical trials, description of the manufacturing facility, proposed product labeling, and post-licensure surveillance plans. The FDA inspects manufacturing sites to evaluate compliance with current Good Manufacturing Practice (cGMP) and other industry standards to ensure purity, potency, safety and consistency in manufacturing. The decision to approve a BLA is based on disease epidemiology, vaccine safety and efficacy 3.

Vaccine Manufacturing

After the approval of a BLA, the licensed vaccine and its manufacturing processes are monitored by the FDA. In addition to the safety tests carried out before licensure, each batch of vaccines, called a lot, undergoes testing for purity, potency, and sterility. Any changes to the vaccine, its label, or manufacturing process must be submitted to the FDA for approval prior to product distribution. Before approving the change, the FDA may require further testing to confirm continued safety of the vaccine 3.


Vaccine Recommendations

Following licensure, the Advisory Committee on Immunization Practices (ACIP) makes recommendations to the Centers for Disease Control and Prevention (CDC) about the optimal use of the vaccine. The ACIP consists of fifteen members who are experts in relevant fields such as immunology, vaccine research and development, internal medicine, pediatric care, nursing, infectious diseases, public health, and social aspects of immunization programs. The committee also includes eight non-voting ex officio members who represent federal government agencies with immunization responsibilities in the United States. Representatives from professional medical associations such as the American Academy of Pediatrics (AAP), the American Academy of Family Physicians (AAFP), the American College of Obstetricians and Gynecologists (ACOG), and the American Medical Association (AMA) participate in these discussions as liaisons, and vaccine recommendations are coordinated or harmonized with AAP, AAFP and ACOG.

Taking into consideration risks, benefits and costs of the vaccine, and disease epidemiology, the ACIP makes a recommendation to CDC including guidelines on which persons should be vaccinated, number of doses, timing, age of administration, and also provides information on adverse events, contraindications and precautions. The ACIP recommendations, once accepted by CDC, may lead to changes in the national immunization schedule which in turn may lead to widespread use of the vaccine. However, the speed of uptake of a new vaccine even after ACIP recommendation can depend on other factors including provider knowledge, costs and financing mechanisms (e.g. if insurance companies cover the cost), and public acceptance 4.

Post-Licensure Safety Surveillance

Clinical trials determine the safety and efficacy of vaccines and the ACIP guidelines are intended to ensure that the benefits outweigh the risks of vaccines. However, clinical trials have limitations. They are usually conducted with strict inclusion criteria, usually in healthy subjects within a certain age group, and therefore results may not be generalizable to all of the populations the vaccine is actually used in after licensure. Trials may not be large enough to detect rare adverse events. For example, if an adverse event occurs at a population baseline rate of 1 in 1,000 people and the vaccine doubles this risk, a clinical trial would need 50,000 subjects (which is larger than most clinical trials) to statistically detect an increased risk posed by the vaccine. Given the near universal use of many vaccines, this adverse reaction could impact 4,000 persons annually, assuming a birth cohort of 4 million and 100% immunization coverage 5. Additionally, due to limited follow-up, delayed onset adverse events may be missed. For these reasons, post-licensure surveillance and studies are vital to ensuring continued vaccine safety.

The Vaccine Adverse Events Reporting System (VAERS) is the national passive surveillance system that is co-administered by the FDA and the CDC. All healthcare providers of childhood vaccines are required to report certain adverse events to VAERS; the public may also report adverse events. Due to the fact that reports are observations from a variety of sources, not physician confirmed events and there are no standardized case definitions, FDA and CDC medical officers review reports of serious events, deaths, and unusual patterns to identify cases that may represent a safety signal that requires additional safety studies. VAERS has the ability to detect signals of rare events. However, the system suffers from both under- (not all adverse events are reported) and over- (non-vaccine related events are reported) reporting. The reports submitted to VAERS may also be incomplete and insufficient to confirm a diagnosis. Furthermore, the system does not have data on the number of vaccine doses administered (doses distributed are a poor estimate of this number) nor the incidence of the adverse event in unvaccinated populations; therefore, it is not possible to calculate incidence rates in vaccinated versus non-vaccinated persons 6. Such comparisons are important to assess whether a given adverse event is causally or coincidentally related to vaccination. For example, if the rates were actually the same, then the data would imply that vaccination did not cause the event. On the other hand, if the rate was higher in vaccinees, this would imply vaccine is playing a causal role. Despite the weaknesses of VAERS for assessing causation, VAERS acts as an early warning system of possible vaccine safety problems that may require further investigation

Vaccine safety signals that arise from clinical trials or passive surveillance such as VAERS are often assessed through large-linked databases which include inpatient, outpatient, laboratory and pharmacy records. These large-linked databases are often referred to as active surveillance as they can be used to rapidly evaluate the safety of newly licensed vaccines or vaccines that are being used in special populations, such as pregnant individuals.

The Vaccine Safety Datalink (VSD) is the active surveillance system which is overseen by the CDC. The VSD is a large-linked database of eight managed care organizations (MCO) that cover 1.8 percent of the United States population under 18 years and 1.5 percent of those at least 18 years 7-9. VSD sites are closed healthcare systems resulting in all clinical records being readily available. These data include details on vaccines, including vaccination dates, vaccine type and lot number. VSD studies often include review of the clinical records, providing far more detail compared to VAERS. Adverse event rates can be calculated using the data, providing the ability to evaluate rare events that could not be detected in clinical trials due to the sample size. VSD data also can be used to assess potential adverse events with delayed onset (as long as the patient is continually enrolled in the site) and among special populations not included in clinical trials for licensure. VSD can rapidly conduct chart reviews in order to validate outcomes, typically using standardized case definitions developed by the Brighton Collaboration 10. The VSD facilitates high quality and rapid vaccine safety studies either in an ongoing basis for pre-specified outcomes using Rapid Cycle Analysis or through ad hoc studies. The VSD allows for calculation of the rates of a given illness in vaccinees versus non-vaccinees, which is often critical in determining if vaccine is playing a causal role. Many VSD studies use self-controlled designs in which the risk of an event is compared in a specified time period after vaccination to another time period among vaccinated persons. This study design using individuals as their own controls adjusts for factors such as health care seeking behavior. Further, VSD can also look at particular time frames after vaccination and compare the rates of the adverse event in that time frame with what would be expected in an unvaccinated population, which can be determined in the VSD.

The Post-Licensure Rapid Immunization Safety Monitoring (PRISM) Network was established to supplement VSD for the 2009-10 H1N1 influenza vaccination program 11. PRISM is now a part of FDA’s post-licensure vaccine safety system. PRISM links 8 state immunization registries to capture who received which vaccines with 4 large health insurance plans capturing some vaccine exposures and a broad range of health outcomes for over 100 million persons 12. PRISM is able to conduct chart review, although not as rapidly as the VSD. PRISM is also used to conduct high quality and rapid vaccine safety studies either in an ongoing basis for pre-specified outcomes using Rapid Cycle Analysis or through ad hoc studies.

Other large-linked databases are used for conducting Rapid Cycle Analysis and ad hoc studies among special population. The Centers for Medicare and Medicaid Services (CMS) focuses primarily on persons at least 65 years of age, including about 14 million vaccines annually. The Department of Defense (DoD) Medical Surveillance System included military personnel and their families which is about 2.6 million persons. The Department of Veterans Affairs (VA) includes military veterans not provided care through DoD and Federal Employees which is about 5 million persons. The Indian Health Services (IHS) includes American Indian and Alaska Native populations, which is about 350,000 persons. CMS, DoD, VA and IHS are able to include chart review and have historically been focused primarily on influenza vaccines.

Coordination of Vaccine Safety Activities

Many government and nongovernmental partners contribute to assessing and monitoring vaccine safety across the United States. The Department of Health and Human Services (HHS)’s National Vaccine Program Office (NVPO) coordinates the efforts of the National Institutes of Health (NIH), the National Vaccine Injury Compensation Program (VICP), FDA, and CDC. The National Vaccine Advisory Committee (NVAC) advises this office in a capacity similar to the ACIP. Other partners such as state and local health departments, academic institutions, professional medical associations, healthcare providers, insurance companies, philanthropic organizations, and vaccine manufacturers, contribute to the provision of safe vaccines and communicate safety risks to the public 1.

Causality Assessment

Causal relationships between vaccines and adverse events can be established by demonstrating an increased risk of the adverse event in vaccine recipients or through documenting the role of a vaccine component in the pathogenesis of the adverse event 13,14. Increased risk is identified from randomized controlled trials or observational epidemiologic studies, such as those performed by the VSD. Prior to licensure, vaccine recipients are compared to placebo or control vaccine recipients in randomized clinical trials to assess differences in adverse events. These randomized studies provide the highest quality of evidence for or against causal associations. However, the sample size in these studies is usually too small to detect small increases in risk and often do not include all the populations who receive the vaccine after licensure.

After licensure, reports to VAERS of adverse events following immunization (AEFI) are monitored. VAERS reports, and case reports in the medical literature of AEFI based on a temporal relationship are commonly misunderstood as causal; but a temporal association alone, even with a hypothesis as to how the vaccine might have caused an adverse event, does not establish a causal relationship 13,14.

If a potential signal for an AEFI is detected, controlled epidemiological studies may be conducted to determine if there is an increased risk in vaccinated persons as compared to people who did not receive the vaccine. Several types of studies have been used including case-control, cohort, and self-controlled studies. These studies are usually not randomized so they can be subject to several possible biases and confounders. A common concern is health care seeking bias. People who chose to be vaccinated may be more likely to seek other healthcare services and consequently be more likely to be diagnosed with other adverse health outcomes. Healthcare seeking bias creates a spurious association between vaccines and adverse events. Usually a single study is insufficient to establish a causal relationship. The evidence is much stronger if consistent associations are found in different studies conducted in different populations by different investigators, sometimes using different methods. Causal relationships have been established between vaccines and adverse events even when the biologic mechanism has not been identified, such as with intussusception following rotavirus vaccines 15-17 and Guillain-Barré syndrome following influenza vaccines 18,19.

Identifying the biologic mechanism for an adverse event, sometimes referred to as mechanistic evidence, can provide strong evidence for a causal relationship even if epidemiologic studies have not documented an increased risk in the general population 13. For example, the yellow fever vaccine virus has been found in liver tissue of patients with viscerotropic disease 20 and measles vaccine viruses have been identified in lung tissue in immunocompromised patients with pneumonia 21,22. Immediate hypersensitivity reactions, including anaphylaxis, usually occur within minutes after exposure. In the absence of other possible exposures, immediate hypersensitivity reactions that occur shortly after a vaccine are usually assumed to have been caused by the vaccine if there have been no other possible exposures to potential allergens. Skin testing with the vaccine can sometimes provide supportive evidence of allergy to specific vaccines or vaccine components, but false positive reactions do occur 23,24.

Individual AEFI should be reviewed in a systematic manner for assessment of possible causal associations. In the U.S., an algorithm approach was developed by the CDC coordinated Clinical Immunization Safety Assessment (CISA) network for review of serious adverse events 14. The method was modified for use in developing countries by the World Health Organization (WHO) 25,26. The method requires persons doing the assessment to collect essential information about the case and to ask key questions in a logical sequence before determining the correct assessment. The validity of the diagnosis should be determined followed by assessment of other possible causes, if a general casual association has been made between the vaccine and the adverse event, and determination if the timing of the event is consistent with existing knowledge. Depending upon the answers to the questions, the algorithm branches to reach a logical conclusion. Use of this approach can prevent the common mistake of assuming causal relationships based upon temporal relationships.

Systematic reviews of causality assessment have historically been conducted in the US by the Institute of Medicine (IOM), now called the National Academy of Medicine (NAM) 13. The IOM used stringent criteria for reviews of the evidence. Starting from a neutral position, the committee weighed the available evidence and published conclusions. However, the majority of the assessments conducted have resulted in conclusions of “The evidence is inadequate to accept or reject a causal relationship”. These assessments are often not helpful and do not reflect the available evidence for or against a causal relationship. The Institute for Vaccine Safety (IVS) of Johns Hopkins uses similarly rigorous criteria for vaccine causality assessment, focusing on epidemiological evidence and potential biological mechanisms. However, determinations are more frequently conclusive than those of the IOM, as the reviews included in this book and on our website ( are intended for clinicians, the public and policy makers who are particularly interested in clarity of messages.

Vaccine Injuries and Compensation

The National Childhood Vaccine Injury Act was passed in 1986 spurred by a major increase in litigation against vaccine manufacturers and healthcare providers who administered vaccines. The primary vaccine of concern at the time was the whole cell pertussis vaccine. This led to vaccine shortages, some manufacturers dropping out of the market, and other problems. Further, it led to the recognition that when a child is vaccinated, that child is not only trying to obtain protection from him/her self but also is protecting the community by preventing spread of the disease to other persons including persons who cannot be protected by vaccines because of contraindications. Society owed individuals injured by vaccines compensation because of the benefits they were trying to bring to society. Thus, the legislation recognized the broad value of vaccines to the population at the expense of a very small number of people who experienced serious vaccine adverse events, and those persons were owed compensation. As a result, it also fostered the need for improving vaccine safety monitoring to help in determining who was eligible for compensation 27. The act authorized the establishment of the National Vaccine Injury Compensation Program (VICP), a no-fault system to provide financial assistance to individuals injured following recommended vaccines incorporated into the program. Compensation is awarded to individuals whose injury is on the vaccine injury table and occurred within the specified time period, or if the claimant can prove the injury was caused by the vaccine. As this causality assessment is of a much lower standard than is used to establish causality scientifically, there are far more compensations than scientifically accepted vaccine injuries. While compensations through the VICP are essential to the national immunization program, they can be interpreted inaccurately by the public as an admission of wrongdoing by the government or vaccine manufacturers 1,28.


In summary, the vaccine safety system is extremely rigorous yet complex and often not fully understood by the public or healthcare providers. Vaccines have an excellent safety record and the system has demonstrated its ability to detect, characterize and prevent rare and serious adverse reactions when they occur. As new vaccines are introduced and populations targeted for vaccines expand, continuous monitoring of vaccine safety is essential for public confidence in immunization programs.


1.         A Comprehensive Review of Federal Vaccine Safety Programs and Public Health Activities (2008).

2.         Food and Drug Administration. Guidance for Industry: Content and Format of Chemistry, Manufacturing and Controls Information and Establishment Description Information for a Vaccine or Related Product. 1999.

3.         Food and Drug Administration. Vaccine Product Approval Process.

4.         Centers for Disease Control and Prevention. Advisory Committee on Immunization Practices (ACIP).

5.         Ellenberg SS. Safety considerations for new vaccine development. Pharmacoepidemiology and drug safety. Aug-Sep 2001;10(5):411-5. doi:10.1002/pds.616

6.         Centers for Disease Control and Prevention (CDC). Vaccine Adverse Event Reporting System (VAERS).

7.         Chen RT, DeStefano F, Davis RL, et al. The Vaccine Safety Datalink: immunization research in health maintenance organizations in the USA. Bulletin of the World Health Organization. 2000;78(2):186-94.

8.         Centers for Disease Control and Prevention. Vaccine Safety Datalink (VSD).

9.         Baggs J, Gee J, Lewis E, et al. The Vaccine Safety Datalink: a model for monitoring immunization safety. Pediatrics. May 2011;127 Suppl 1:S45-53. doi:10.1542/peds.2010-1722H

10.       Kohl KS, Bonhoeffer J, Braun MM, et al. Advances in Patient SafetyThe Brighton Collaboration: Creating a Global Standard for Case Definitions (and Guidelines) for Adverse Events Following Immunization. In: Henriksen K, Battles JB, Marks ES, Lewin DI, eds. Advances in Patient Safety: From Research to Implementation (Volume 2: Concepts and Methodology). Agency for Healthcare Research and Quality (US); 2005.

11.       Salmon D, Yih WK, Lee G, et al. Success of program linking data sources to monitor H1N1 vaccine safety points to potential for even broader safety surveillance. Health affairs (Project Hope). Nov 2012;31(11):2518-27. doi:10.1377/hlthaff.2012.0104

12.       Baker MA, Nguyen M, Cole DV, Lee GM, Lieu TA. Post-licensure rapid immunization safety monitoring program (PRISM) data characterization. Vaccine. Dec 30 2013;31 Suppl 10:K98-112. doi:10.1016/j.vaccine.2013.04.088

13.       Institute of Medicine. In: Stratton K, Ford A, Rusch E, Clayton EW, eds. Adverse Effects of Vaccines: Evidence and Causality. National Academies Press (US); 2012.

14.       Halsey NA, Edwards KM, Dekker CL, et al. Algorithm to assess causality after individual adverse events following immunizations. Vaccine. Aug 24 2012;30(39):5791-8. doi:10.1016/j.vaccine.2012.04.005

15.       Aliabadi N, Tate JE, Parashar UD. Potential safety issues and other factors that may affect the introduction and uptake of rotavirus vaccines. Clinical microbiology and infection : the official publication of the European Society of Clinical Microbiology and Infectious Diseases. Dec 01 2016;22 Suppl 5:S128-s135. doi:10.1016/j.cmi.2016.03.007

16.       Yih WK, Lieu TA, Kulldorff M, et al. Intussusception risk after rotavirus vaccination in U.S. infants. The New England journal of medicine. Feb 06 2014;370(6):503-12. doi:10.1056/NEJMoa1303164

17.       Parashar UD, Cortese MM, Payne DC, Lopman B, Yen C, Tate JE. Value of post-licensure data on benefits and risks of vaccination to inform vaccine policy: The example of rotavirus vaccines. Vaccine. Jun 27 2015;doi:10.1016/j.vaccine.2015.05.094

18.       Halsey NA, Talaat KR, Greenbaum A, et al. The safety of influenza vaccines in children: An Institute for Vaccine Safety white paper. Vaccine. Dec 30 2015;33 Suppl 5:F1-f67. doi:10.1016/j.vaccine.2015.10.080

19.       Salmon DA, Proschan M, Forshee R, et al. Association between Guillain-Barre syndrome and influenza A (H1N1) 2009 monovalent inactivated vaccines in the USA: a meta-analysis. Lancet. Apr 27 2013;381(9876):1461-8. doi:10.1016/s0140-6736(12)62189-8

20.       Seligman SJ. Risk groups for yellow fever vaccine-associated viscerotropic disease (YEL-AVD). Vaccine. Oct 07 2014;32(44):5769-75. doi:10.1016/j.vaccine.2014.08.051

21.       Measles pneumonitis following measles-mumps-rubella vaccination of a patient with HIV infection, 1993. MMWR Morbidity and mortality weekly report. Jul 19 1996;45(28):603-6.

22.       Angel JB, Walpita P, Lerch RA, et al. Vaccine-associated measles pneumonitis in an adult with AIDS. Ann Intern Med. Jul 15 1998;129(2):104-6.

23.       Wood RA, Berger M, Dreskin SC, et al. An algorithm for treatment of patients with hypersensitivity reactions after vaccines. Pediatrics. Sep 2008;122(3):e771-7. doi:10.1542/peds.2008-1002

24.       Wood RA, Setse R, Halsey N. Irritant skin test reactions to common vaccines. The Journal of allergy and clinical immunology. Aug 2007;120(2):478-81. doi:10.1016/j.jaci.2007.04.035

25.       Tozzi AE, Asturias EJ, Balakrishnan MR, Halsey NA, Law B, Zuber PL. Assessment of causality of individual adverse events following immunization (AEFI): a WHO tool for global use. Vaccine. Oct 17 2013;31(44):5041-6. doi:10.1016/j.vaccine.2013.08.087

26.       World Health Organization. Causality assessment of an adverse event following immunization (‎AEFI)‎: user manual for the revised WHO classification, 2nd ed. World Health Organization; 2018.

27.       National Childhood Vaccine Injury Act, (1986).

28.       Institute of Medicine (US) Committee on Review of Priorities in the National Vaccine Plan. The Safety of Vaccines and Vaccination Practices. Priorities for the National Vaccine Plan. National Academies Press (US); 2010.

 [MD1]Note to publisher to add hyperlink to online version:

Note to publisher to add hyperlink to online version:

 [MD3]Note to publisher to add hyperlink to online version: