Skip Navigation

Center of Excellence in Regulatory Science and Innovation


As one of the world’s leading institutions in higher education, Johns Hopkins offers an ideal location for the development of fundamental new knowledge regarding regulatory science.

Our CERSI’s scholarly activities span the University’s schools, divisions, institutes and centers and have the singular focus of enhancing the FDA’s evidentiary basis for decision-making. Our research activities are focused on numerous priority areas for the agency, including: improving clinical studies and evaluation; strengthening the social and behavioral sciences to support informed decisions; and innovating the use of real-world evidence (RWE) in the life-cycle evaluation of FDA regulated products. 

Several active areas of investigation are described in detail below, although there are many others, ranging from understanding the effects of e-cigarette advertising on consumer intentions to use to elucidating the mechanism of neurotoxic side-effects of cancer immunotherapies, that leading scientists at Johns Hopkins are pursuing in order to strengthen the FDA’s ability to fulfill its mission through enhanced regulatory science and innovation.

Jump to Project: 

Current Research




Digital Health


Patient Preferences and Patient Reported Outcomes (PROs)

Real World Evidence

Tobacco Products


Prior Research


Current Research

Improving precision and power in randomized trials for COVID-19 treatments using covariate adjustment, for ordinal or time to event outcomes

It is important that the FDA and manufacturers agree upon the endpoints and statistical analyses to support the effectiveness of products under investigation for treating COVID-19 patients. The aim of this study is to help inform the choice of endpoints and analysis methods to be used in future COVID-19 treatment trials. The study was informed by the work of a World Health Organization (WHO) working group and other researchers that identified core outcomes of importance when caring for patients with COVID-19, including outcomes observed during acute hospitalization, as well as after discharge. This study builds on that previous work to recommend specific estimands for those core outcomes as well as the appropriate statistical methods to analyze them. To evaluate the performance of the estimators that they present, the team simulated two-arm, randomized trials comparing a hypothetical COVID-19 treatment versus standard of care, where the primary outcome is ordinal or time-to-event, using real-world data from hospitalized patients. This work will assist FDA and manufacturers at a critically important juncture to best ascertain the effectiveness of candidate therapies to treat COVID-19. 

Project team leads: Jodi Segal, MD, MPH and Michael Rosenblum, PhD (Johns Hopkins)

For more information, please contact Jodi Segal at

Back to top

Assessing real world use of pharmaceuticals among pregnant and lactating women

Many women need to take medications to treat medical conditions during pregnancy and lactation. The FDA has established guidelines and recommendations intended to maximize drug safety and efficacy in these women while minimizing toxicity to the fetus. To best address the health care needs of these women, it is important to monitor how well FDA guidelines are being followed and whether any changes to current dosing schedules need to be made for this special population given the dramatic physiology changes that occur during pregnancy which can alter drug absorption, bioavailability, distribution, and clearance. However, little is known about the distribution of medication use pregnant and lactating women or how it compares to recommendation for use in this special population of women. In this project, we address this critical knowledge gap by assessing medication and supplement use among pregnant and lactating women using “real world” data that was collected over the past two decades from a variety of sources including claims, electronic medical record, and self-report databases. We will evaluate women’s frequency of use - prior to, during, and after pregnancy – including for subgroups of these women (e.g. obese, low socioeconomic status, advanced maternal age) and will also identify constellations of medications/supplements being used in tandem. Use patterns will be compared to FDA guidelines and recommendations, including changes over time. Our results will provide critical knowledge regarding how women’s use compares to FDA guidelines and labeling changes. Additionally, they can inform future to studies that seek to optimize dosing schedules in pregnant and lactating women, by aiding drug prioritization and medication/supplement combinations to be tested.

Project PI: Christine Ladd-Acosta, PhD (Johns Hopkins)

For more information about the project, email Christine Ladd-Acosta at

Back to top

Understanding flavors in Electronic Nicotine Delivery Systems (ENDS) (Phase 2)

This work represents an extension of earlier CERSI funded work (Kennedy RD et al). The FDA has the authority to regulate and restrict the marketing of tobacco products. However, given the recency of ENDS products, limited research exists to inform the regulation of their marketing. The earlier study identified useful data sources and compelling findings to warrant ongoing surveillance and continued analysis of the content of ENDS advertising to understand how design features including flavors are presented in advertising content. In the new study, the team will maintain a clear focus on understanding the role flavors play in advertisements. The team plans to broaden their focus to also examine how nicotine, its concentration/volume/mass and type (nicotine or nicotine salts for example) are depicted/communicated in ads. The team will also assess the presence and content of health warnings now required in ENDS advertisements. Findings of this study can be used to inform FDA rulemaking regarding the marketing of ENDS and to guide other public health agencies policies and messaging regarding the role of flavors in ENDS.

Project team lead: Ryan David Kennedy, PhD (Johns Hopkins)

For more information, please contact Ryan David Kennedy at

Back to top

Using social media for tobacco regulatory intelligence

Effective tobacco control depends on timely, accurate surveillance.

However, existing strategies rely mostly on cumbersome surveys that cannot be cost-effectively scaled to yield timely insights. To fill this gap, experts are increasingly turning to social media surveillance. Our goal is the advancement of tobacco regulatory science through the application of artificial intelligence (AI) methods to social media data. We will use these methods to discover new tobacco products mentioned on social media but as yet unknown to the FDA's Center for Tobacco Products. Our analysis will apply AI algorithms driven by natural language processing (NLP) to millions of messages collected from Twitter and Reddit.

Project Team Leads: Mark Dredze, PhD (Johns Hopkins University), John W Ayers, PhD (University of California- San Diego)

For more information, please contact Mark Dredze at

Back to top

Active surveillance of medical device safety and outcomes using EHRs:  prostate cancer partial gland ablation technologies 

For this project, researchers at Johns Hopkins University Medical School and Weill Cornell Medical College will develop natural language processing (NLP) tools to perform active surveillance of medical devices used to treat prostate cancer with partial gland ablation. While traditionally, prostate cancer has been treated with whole gland treatments such as radical prostatectomy or gland radiation therapy, there is growing interest in partial gland ablation which aims to decrease the side effects of prostate cancer treatment including erectile dysfunction and urinary incontinence. In light of the growing interest in partial gland ablation, there exists a need to perform active surveillance of the safety, side-effect profile, and cancer control outcomes with the various available medical devices use for prostate tissue ablation such as high intensity focused ultrasound (HIFU) and cryoablation.  NLP offers a cost-effective approach for monitoring electronic health records (EHRs) for adverse events and cancer-specific outcomes. The NLP tools developed for this project may be applicable to other medical devices and disease processes.

Project Team Leads: Christian Pavlovich, MD (Johns Hopkins), Jim Hu, MD (Cornell University) (Overall scientific lead), Art Sedrakyan, MD, PhD (Cornell University)

For more information, please contact Christian Pavlovich at

Back to top

Assessing Physiological, Neural and Self-Reported Response to Tobacco Education Messages  

This project will use neuroimaging, physiological and self-report measures to assess response to FDA tobacco education messages. Tobacco education campaigns are an important component of tobacco control and have contributed to significant decreases in US smoking rates. Developing effective tobacco education campaigns requires testing message strategies, to ensure that campaigns use the most effective messages for different target populations. This type of testing typically asks individuals to self-report their response to the message (e.g., whether the message would make them less likely to smoke cigarettes). While these self-report measures are useful, other methods, such as neuroimaging and assessing physiological response to the messages, can be a useful and innovative supplement. Physiological and neural response can be measured using indicators such as heart rate variability (HRV), galvanic skin response (GSR), and facial electromyography (EMG), which can assess arousal and affective response to messages, while tools such as eye tracking and neuroimaging can measure attention and levels of activation in key areas in the brain associated with message processing and message acceptance. Research indicates that often, these techniques are more effective than self-report measures at predicting ‘real world’ message effectiveness. This project will specifically examine response to messages in the Fresh Empire campaign, Real Cost e-cigarette campaign, and This Free Life campaign, among a sample of adolescents and young adults.  Ultimately, this work will contribute to FDA understanding of effective tobacco education messaging tactics and will identify mechanisms underlying such message effects.

Project Team Lead: Meghan Brigid Moran (Johns Hopkins)

For more information, please contact Meghan Moran at

Back to top

A systematic review of existing patient-reported outcome instruments for glaucoma

In this project, the investigators will conduct a systematic review in order to identify and assess the quality of all existing patient-reported outcome (PRO) instruments relevant to patients with glaucoma that have been described in the published medical literature. This is important because in order to effectively capture patient experiences with emerging therapeutic modalities such as minimally-invasive glaucoma surgery (MIGS), we need to use well-developed, appropriately validated patient-reported outcome measures suitable for the relevant patients and clinical questions. In the case of MIGS, those are patients with mild to moderate glaucoma, and an ideal PRO measure needs to be responsive to changes along a peri-operative timeline (e.g. both short-term post-operative experiences and longer-term changes corresponding to surgical outcomes). An important step towards this goal is determining whether such measures currently exist in the published medical literature. This project will focus specifically on characterizing the development and validation methodology for these instruments. Goals of the systematic review are; firstly, identification of existing, published PRO instruments which could be sensitive to patient-important outcomes of minimally-invasive glaucoma surgery; secondly, identification of areas in which existing instruments are sub-optimal for any of the following applications: a) capturing experiences of patients with mild glaucoma who retain good vision, b) capturing the experience of surgery, which includes both short-term (peri-operative) and long-term changes; and thirdly, identification of glaucoma-related PROs which were explicitly designed or validated using modern psychometric methods. The results of this review will help FDA to improve the clinical and postmarket evaluation of MIGS and other novel therapeutic modalities for glaucoma patients.

Project Team Lead: Amanda Kiely Bicket, MD (Johns Hopkins)

For more information, please contact Amanda Bicket at

Back to top

Histology, pathology and histopathology of humanized mice

The goal of this project is to create a knowledge base for the normal histology and histopathological responses to drugs in chimeric (part human/part mouse) mice that can be used to support qualification of the model as a drug development tool and adoption of the model by the drug development industry as such. This is important because the FDA is developing mouse models that have human components (chimeric), and therefore provide responses to drug toxicities, immune reactions, and metabolism issues that are more predictive than any existing preclinical models. The goal of improved predictively is to minimize serious unanticipated adverse reactions in early clinical trials that can injure and upon occasion result in death of healthy human volunteers. In addition, the models may also lead to early identification of adverse events that are currently only recognized after a drug has been marketed and many people exposed. Although chimeric mice have been created for a number of years, the degree of human components, the complexity, and the utility of the models have expanded rapidly. Greater understanding of the chimeric mouse model will be accomplished through the compilation of a histology, pathology, and histopathology atlas for these models. This collaborative effort will provide consistent expert pathology/histopathology support for ongoing FDA-based humanized mouse experiments. This project would help in the development and adoption of these models for use in the drug development industry and provide FDA reviewers a solid and well defined knowledge base to inform future regulatory decisions when these models have been used.

Project Team Lead: Kathleen Gabrielson, DVM, PhD (Johns Hopkins)

For more information, please contact Kathleen Gabrielson at

Back to top

Evaluating bulk drug substances used to compound drugs for patients with ASD

For this project, Johns Hopkins University CERSI researchers will develop guidelines to systematically study available safety and effectiveness information on certain bulk drug substances that are used to compound drugs for patients with autism spectrum disorder (ASD). The FDA intends to consider this information as it evaluates these substances for use in compounding. This is important because drugs compounded by outsourcing facilities are exempt from certain sections of the Federal Food, Drug, and Cosmetic Act (FD&C Act), including requirements for FDA approval of a drug and labeling with adequate directions for use, if they meet certain conditions described under section 503B of the FD&C Act. One of the conditions is that if the outsourcing facility compounds drugs using bulk drug substances (active pharmaceutical ingredients), then the substances must be used to compound drugs in shortage or appear on the 503B bulks list established by the FDA identifying bulk drug substances for which there is a clinical need. To determine whether to include substances nominated by the public on the 503B bulks list, FDA intends to evaluate several factors, including, in certain cases, the safety, any evidence of effectiveness, and the historical and current clinical use of drugs compounded using each substance. 

Project Team Leads: Karen Robinson, PhD and Heather Volk, PhD (Johns Hopkins)

For more information, please contact Karen Robinson at

Project Final Report (PDF)

Back to top

Investigating how flavors influence consumer decision-making about ENDS usage

The objectives of the proposed project are to examine how different electronic nicotine delivery systems (ENDS) flavors and advertising tactics are associated with (a) consumer perceptions of product qualities and (b) intentions to use the product. This is important because ENDS, also called electronic cigarettes, e-cigarettes, and vaporizers, are only newly regulated by the FDA, and given the recent introduction of ENDS products to the market, limited research exists to inform the regulation of their marketing. Flavors are a unique and important aspect of ENDS. Flavors are commonly reported by both youth and adult ENDS users as a main reason why they use ENDS. Flavors are a regulatory area of interest, and the FDA has issued an advance notice of proposed rulemaking (ANPRM) “to obtain information related to the role that flavors play in tobacco products,” with a specific interest in how flavors spur youth product initiation. The proposed project will accomplish its stated aims in two phases by obtaining youth’s product perceptions and use intentions in response to different advertising tactics. This project employs a mixed experimental design in which participants will view ENDS ads with key features (e.g., the presence of color) present or absent. This approach will provide evidence as to how specific advertising features used to convey ENDS flavors affect product perceptions and intentions to use ENDS.

Project Team Lead: Meghan Brigid Moran, PhD (Johns Hopkins)

For more information, please contact Meghan Brigid Moran at

Back to top

Positron Emission Topography (PET) to evaluate CNS response to immunotherapy

The project’s goal is to generate fundamental new knowledge regarding the causes of serious adverse drug events associated with cellular immunotherapies, including cytokine release syndrome and neurotoxicity. Lymphocytes, a type of white blood cell in the body’s immune system that are engineered to express chimeric antigen receptors (CAR-T cells), have shown remarkable antitumor effects in hematological malignancies that have led to two FDA approvals. However, even the most successful CAR-T cell therapies are not accomplished without toxic side effects in humans. Cytokine release syndrome is a common complication that can be managed clinically. Neurotoxicity is observed in a subset of patients, and the pathogenicity remains to be elucidated. Although distribution and persistence of CAR-T cells in blood and tumors can be monitored through cell and tissue-based technologies, their disposition in various tissues is poorly understood. Positron emission tomography (PET) is a quantitative technique that could be used to quantify the CAR-T cell distribution in various tissues non-invasively and without compromising cell specificity. The cytokine IL-13 binding IL-13Ra2 receptor is highly expressed in high-grade glioblastoma, and contributes to chemoresistance to temozolamide therapy. Furthermore, IL- 13Ra2 expression correlates with immune suppressive tumor microenvironment in glioblastoma, thus providing rationale for further characterization of IL-13Ra2 CAR-T cell trafficking and distribution in tumors. Aims include (1) to evaluate trafficking and persistence of IL-13Ra2 CAR-T cells in vivo and relate them with neurological and systemic effects, (2) to assess the intratumoral penetration and distribution of IL-13Ra2 CAR-T cells in solid tumors in vivo by PET and ex vivo by autoradiography and immunohistochemistry (or fluorescence microscopy,) and (3) to analyze the therapeutic efficacy of systemic vs. local stereotactically injected IL-13Ra2 CART cells on tumor growth to correlate with neurological and systemic effects. 

Project Team Lead: Sridhar Nimmagadda, PhD (Johns Hopkins)

For more information, please contact Sridhar Nimmagadda at

Back to top

Synthesizing Real-World Data For Regulatory Decision Making In Single-Group Medical Device Clinical Studies

The goal of this project is to synthesize real-world evidence (RWE) for selecting non-concurrent controls, establishing performance goals, and widely disseminating relevant methods and software. This is important because the FDA is mandated to consider “the least burdensome appropriate means of evaluating device effectiveness that would have a reasonable likelihood of resulting in approval.” Thus, the FDA allows for single-group studies for pre-market device evaluation when “the device technology is well developed and the disease of interest is well understood.” These studies may compare a device to non-concurrent controls or performance goals derived from non-concurrent information. However, there are challenges to such approaches, some of which may be addressed by RWE. We will develop standards to quantify the quality, relevancy, reliability, comprehensiveness, and completeness of an RWE source. We will also extend the use of propensity scores to this setting and develop statistical methods for selecting non-concurrent controls or choosing performance goals for single-group device studies.  We will develop recommendations for conducting and objectively reporting sensitivity analyses and develop software for propensity score-based methods for selecting non-concurrent controls or performance goals. In addition, we will generate case studies, short-courses and other materials to illustrate these approaches to regulators, sponsors and other stakeholders.

Project team lead: Chenguang Wang, PhD (Johns Hopkins)

For more information, please contact Chenguang Wang at

Back to top

Assessing the Robustness of Clinical Machine Learning Models to Changes in Context of Use

The performance and reliability of machine learning (ML) models are known to be dependent on the "context" (i.e., aspects of the training dataset) for which the model was trained to be used. In this project, we want to develop a framework for measuring robustness of machine learning models to shifts in context that commonly occur in real-world data. This framework will help regulators assess the risk associated with a particular intended use through the context shifts that can occur, and the robustness tests we develop can serve as example tests that model developers can run to provide further evidence that their devices are robust to changes in contexts that are likely to occur.

Project team lead: Suchi Saria, PhD (Johns Hopkins)

For more information, please contact Suchi Saria at

Back to top

Improving Real World Measurement of Ambulatory Adverse Drug Events (ADEs) in Minority and Special Need Populations Using EHR Records 

Adverse drug events (ADEs) are costly but largely preventable. The National Action Plan for Adverse Drug Event Prevention (NAPADEP) recognizes that several patient populations may be especially vulnerable to ADEs. Methodological advances in natural language processing (NLP) and machine learning (ML) has enabled researchers to identify ADEs in real world data sources such as EHRs (e.g., clinical notes; a.k.a. EHR’s free-text) in addition to traditional data sources such as claims. We are proposing a CERSI-funded project to measure the ADE disparities of type 2 diabetes (T2D) among African American patients of the Johns Hopkins outpatient delivery system. Results of this research will span across advancing regulatory sciences, disseminating results/knowledge, and informing regulatory decision making hence having the potential to advance public health outcomes.

Project team lead: Jonathan Weiner, PhD and Hadi Kharrazi, PhD (Johns Hopkins)

For more information, please contact Jonathan Weiner at

Back to top

The Regulatory Effect of Substrate Mechanical Properties on the Production and Immunomodulation Efficacy of Mesenchymal Stem Cells  

Mesenchymal stem (stromal) cells (MSCs) are being widely studied for regenerative medicine and immune regulation applications. For example, MSC transplantation represents a promising therapeutic intervention for inflammatory diseases, and there are currently over 200 clinical trials investigating the immunomodulation function of MSCs for the treatment of Graft versus host disease (GVHD), chronic heart failure, Crohn’s disease, etc. However, MSCs are heterogeneous and have multi-modal activity, which is a major challenge for MSC product development and quality control. In recent review of MSC products (e.g. remestemcel-L meeting), FDA has identified that MSC quality attributes or characteristics are often not very well controlled and that the cell manufacturing process could substantially affect MSCs’ clinical effectiveness. The overarching goal of the collaboration  is to better understand and improve MSC manufacturing. This collaboration will both improve our understanding of fundamental MSC biology and advance regulatory science in manufacturing MSC therapeutics.

Project team lead: Luo Gu, PhD (Johns Hopkins)

For more information, please contact Luo Gu at

Back to top


Prior Research

Back to top

Evaluating Development Strategies and Regulatory Outcomes for FDA-Approved Biologics

The goal of this project is to analyze publicly available information derived from the Summary Basis of Approvals (SBOAs) to characterize the clinical development programs of approximately 75 Biologics approved during the past decade. We will identify product and clinical development characteristics associated with Phase 3 success, as determined by the number of studies needed to accrue the necessary evidence for approval and having a first-cycle approval.  Our analyses are motivated by the fact that several key pieces of information (e.g., maximal tolerated dose, evaluation of dose in well-designed Phase 2 studies) facilitate the planning, conduct and interpretation of registrational Phase 3 studies. However, there are important differences between small molecule and biologic products that have implications for clinical development.  Thus, we will focus on development programs such as the studies used to develop the rationale for the dosing regimen, as well as the molecular and therapeutic features of the product. We are also interested in how efficiencies such as adaptive designs are used, whether their use is associated with product features (e.g., indication, molecule type), and what effect this has on the timing and program approval. By highlighting various pathways used for development, as well as efficiencies and outcomes, our findings will generate fundamental new knowledge of high interest and relevance to sponsors and regulators alike. Our team includes a diverse group of experts at CDER and Johns Hopkins with expertise in clinical development, regulatory policy, clinical medicine, pharmacoepidemiology and biostatistics.

Project team lead: G. Caleb Alexander, MD, MS (Johns Hopkins)

For more information, please contact G. Caleb Alexander at

Increasing compliance with registration and reporting requirements

The Food and Drug Administration Amendments Act of 2007 (FDAAA) requires that clinical trials of regulated products are registered in advance and that their results are posted on, a database maintained by the National Library of Medicine (NLM). These requirements are designed to increase research transparency, improve access to postmarket safety information, and improve the information about the effects of regulated products for underrepresented groups.  Compliance with these requirements is poor, especially by non-industry sponsors, and compliance may remain low if academic medical centers (AMCs) and other non-industry sponsors do not implement appropriate policies and procedures to increase compliance. 

The goal of this study is to identify best practices for increasing compliance and to determine if those practices are being followed. First, we will conduct a survey Protocol Registration and Results System (PRS) of administrators and describe the policies and procedures they use for registering and reporting clinical trials.  We will quantify the resources that institutions allocate to supporting compliance. For our second aim, we will identify the proportion of trials that are compliant with reporting requirements at each institution. We will describe the policies, procedures, and resources at the best performing institutions, and we will interview administrators to identify best practices for achieving compliance.

Project team lead: Evan Mayo-Wilson, DPhil (Johns Hopkins)

For more information, please contact Evan Mayo-Wilson at

Understanding Flavors in Electronic Nicotine Delivery Systems 

The goal of this project is to use e-cigarette industry documents including patents to identify flavorants added to e-liquids and any details relevant to public health associated with flavor ingredients including toxicity and addiction. We will also analyze print media advertisement content and associated audiences for e-cigarette products to understand how flavors are portrayed and which sub-populations see which products. This is important because in 2007, a new class of products called ENDS entered the US marketplace. ENDS include a diverse set of products that use a liquid, generally propylene glycol and/or vegetable glycerin, that contains nicotine as well as varying compositions of flavorings and other ingredients which are heated into an inhalable aerosol. Lab-based studies have identified that irritant levels from ENDS aerosols are sufficient to produce respiratory irritation. We will examine e-liquid flavorant constituents to identify what ingredients are being used to imbue specific flavor qualities to ENDS aerosol.  These analyses, which will be conducted using advanced evidence synthesis methods and comprehensive information regarding advertisement content and exposure, will yield important new information to the Center for Tobacco Products as regulations governing e-cigarettes and related products are developed and refined.

Project team lead: Ryan David Kennedy, PhD (Johns Hopkins)

For more information, please contact Ryan David Kennedy at

Evaluating Drug Development in Pediatrics

The overarching goal of this project is to improve the success rate and accelerate pediatric drug development through the retrospective analysis of relevant existing adult and pediatric clinical trial data. This is important because despite the obvious need for safe and effective treatments for children, pediatric drug development lags behind drug development in many ways. To accomplish our goal, we will partner with the FDA to develop Bayesian methods to aid inference and modeling in pediatric drug development. This inferential paradigm provides a structured way to combine information from disparate sources within the framework of probability theory. If a sponsor wishes to show safety and efficacy for a pediatric indication of a compound already approved for adults, it would be more efficient to leverage that information in the drug design, especially if the data includes pharmacokinetics and pharmacodynamics. With this data from adult subjects, one can simulate clinical trials and predict outcomes, along with associated uncertainties, under various dosing and alternative schedules. Careful incorporation of data from outside the pediatric study in the design and analysis of that study will improve efficiency and ultimately, the Administration's ability to better extrapolate methods and best practices in trial design. 

Project team lead: Gary Rosner, ScD (Johns Hopkins)

For more information, please contact Gary Rosner at

Visualizing post-market multi-source information of FDA-regulated products

The goal of this project is to develop methodology for efficiently aggregating, integrating, and visualizing adverse event (AE) information from multiple sources; accurately classifying reports for causality; and, successfully identifying safety patterns in the FDA Adverse Event Reporting System (FAERS). The FDA receives nearly 2 million reports of AEs occurring after the use of drugs and therapeutic biologics via the FDA Adverse Event Reporting System (FAERS) each year, and they use FAERS primarily to identify “signals” of possible adverse drug effects. Historically, identifying these “signals” of possible adverse drug effects has required review by human experts. Experts further evaluate the product-related AE information in other sources, such as published articles. With the increasing number of reports and the size of external information, it has become impossible for human experts to review everything thoroughly. FDA has already implemented computer-based approaches to improve the efficiency and scientific rigor of the human expert review processes. However, definitive and high-performing solutions have yet to be found. This project will focus on three primary goals. The first goal is to develop a system for aggregating data from multiple sources and generating compelling visualizations for the experts. Second, the team will evaluate a combination of Natural Language Processing (NLP) and unsupervised learning for classifying FAERS reports as to their likely value for causal inference. Third, the team will apply an existing text processing tool to FAERS narratives. If successful, this project will enhance human expert review of FAERS reports and other external data. It will also demonstrate a new approach for case classification that does not require large numbers of classified cases and incorporates human expert knowledge into the algorithm.

Project Team Lead: Taxiarchis Botsis, MSc, PhD (Johns Hopkins)

For more information, please contact Taxiarchis Botsis at

Evaluating SGLT-2 inhibitors and chronic kidney disease in patients with diabetes

More than 30 million Americans have diabetes (approximately 1 in 10) and 90% to 95% of them have Type II Diabetes (T2D). T2D leads to macrovascular and microvascular damage, including 2 times the risk of cardiovascular disease (CVD) and 2 times the risk of chronic kidney disease (CKD) compared to the general population. Sodium-glucose cotransporter-2 (SGLT-2) inhibitors block the reabsorption of glucose in the kidney, increase glucose excretion, and lower blood glucose levels. They are indicated as an adjuvant therapy to diet and exercise for improving glycemic control in T2D. There are three SGLT-2 inhibitors that are currently approved by the US Food and Drug Administration (FDA): empagliflozin, canagliflozin, and dapagliflozin. This project seeks to assess whether the established cardiovascular disease (CVD) benefit for SGLT-2 inhibitors in patients with Type II Diabetes extends to those with lower estimated glomerular filtration rate (eGFR) as well as to establish whether the known risk of a urinary track infection (UTI) is different for those patients. It will then assess whether these risks are different across the CVD stage. Finally, the project will assess the effectiveness of SGLT-2 inhibitors on the progression of diabetic kidney disease.

Project Team Lead: Jung-Im Shin, MD (Johns Hopkins)

For more information, please contact Jung-Im Shin at

Patient-Focused Drug Development For Glaucoma

Recent innovation in glaucoma procedures has led to the development of minimally invasive glaucoma surgical (MIGS) devices, which are claimed to be safer than other glaucoma surgical procedures and to reduce the need to use eye drops.  Prior evaluation of these devices has not incorporated patient preference information.  This project will involve glaucoma patients, their caregivers, the physicians, researchers and regulators (i.e. the FDA) to define the essential characteristics for a framework to identify outcomes for MIGS clinical trials that are of direct importance to patients.  Aims include (1) conducting in-depth interviews with patients to further understand their perspective; and (2) developing and pretesting a survey instrument for subsequent quantitative preference studies.

Patient preference information provides insight on the relative desirability and acceptability of the benefits and risks of therapies in managing medical conditions, yet are not well represented as outcomes in current studies of MIGS devices.  Next steps involve engaging with patients with mild to moderate glaucoma through a combination of qualitative and quantitative approaches using semi-structured interviews and surveys to under their preferences.  This CERSI project will contribute to the rapidly developing methodological area of patient preference research and how patient reported outcomes can be incorporated into the regulatory decision making.

Project team lead: Tianjing Li, MD, PhD (Johns Hopkins)

For more information, please contact Tianjing Li at

Comparing Qualitative and Quantitative Approaches to Eliciting Patient Preferences: A case study on innovative upper limb prostheses

Emerging prosthetic technologies such as novel robotic limbs, implantable EMG electrodes, new surgical techniques for prosthetic integration, and electrical stimulation for sensory feedback are promising solutions to some of the challenges faced by amputees. However, they pose additional risks, such as the possibilities of poor surgical outcomes, implanted device malfunction, or infection. These devices are inherently preference-sensitive, as patients work closely with clinicians to choose a prosthesis that fits their needs, and on a daily basis decide whether to use the device. This study is timely, as several IDEs and an HDE are currently under review or approved, a variety of clinical trials are underway and/or planned in the next 5 years, and only one de novo has been granted in this space (for a device that did not involve implanted components).

Approaches to collecting patient preference information are varied, and different methods are applicable to different scenarios. This project will assess patient views of novel upper limb prosthesis benefits and risks using several different preference elicitation approaches, including interviews, focus groups, and surveys. As part of this project, a one day workshop with patients with upper limb amputations and FDA reviewers will be conducted. At this workshop patients and FDA reviewers will complete prioritization exercises and the results of these exercises will be discussed. This work will contribute to the developing science of patient input by helping to identify the advantages and challenges associated with different preference-elicitation approaches. In addition, it will serve to familiarize FDA staff with patient-preference methods and identify similarities and differences in benefit-risk attitudes between FDA reviewers and patients.

Project team lead: John F P Bridges, PhD (Johns Hopkins)

For more information, please contact John Bridges at

use of existing drugs as part of novel treatment strategies in low resource settings

The ultimate goal of Project CURE (Collaborative Use Repurposing Engine) is to obtain information on the use of novel treatment strategies using existing drugs  directly from clinicians and to organize it in a way to encourage the formal study by institutions involved in drug development. By doing so, ineffective or harmful uses may be avoided. The program, which reflects an FDA-NIH collaboration, focuses on drugs for infectious diseases lacking adequate treatments including neglected tropical diseases, emerging infectious threats and infections caused by antimicrobial-resistant organisms. The program will involve four major tasks: needs assessment; application specification; development of application; and beta testing. As part of this program, the Johns Hopkins CERSI will partner with a mobile technology company, emocha Mobile Health Inc, to develop and pilot an application to assist in the implementation and dissemination of Project CURE.

The Collaborative Use Repurposing Engine project (Project CURE) is a partnership between the U.S. Federal Drug Administration (FDA) and the U.S. National Institutes of Health (NIH).  Project CURE is a collaboration to build an internet-based repository which will allow clinicians around the world to report novel uses of existing drugs to treat drug-resistant (DR) infectious and neglected tropical diseases through a mobile device application.  The repository will capture clinical outcomes when drugs are used for new indications, in new doses or new combinations.  The FDA intends the future Project CURE application to serve as a dynamic platform and resource to exchange experience with difficult-to-treat infections.  Ultimately, the Project CURE repository could enrich the drug armamentarium to treat infectious disease, especially in settings where DR infectious and neglected tropical diseases occur more frequently.

Project team lead: John F P Bridges, PhD (Johns Hopkins)

For more information, please contact John Bridges at

Bayesian Approaches for Meta-Analyses of Randomized Clinical Trials When Analyzing Data for Rare Adverse Effects

Meta-analysis has an important role in the evaluation of the safety of medical products, whether they are drugs, biologics or medical devices. Meta-analysis may be used to pool information on events over trials, conducted prior to the marketing or after the marketing of a product. However, with rare events, traditional meta-analysis methods may be ill-defined or have poor performance properties. One potential area for the meta-analysis of safety is in the area of Integrated Summaries of Safety (ISS), which the FDA requires companies to provide as part of submitted marketing applications for drugs and biologics. Another area in which a meta-analysis of rare events occurs is when a specified safety issue is explored within a class of products. However, for many products, rare but serious events may be infrequent and some trials may not have events in one or both of the treatment arms, the latter being called “zero-event trials”. These situations create methodological and interpretative issues. This project will examine the robustness of Bayesian approaches to deal with the issue of rare events including zero event trials. The first step of this undertaking is the design and execution of a comprehensive literature review and gap analysis focused on Baysian and frequentist analytic approaches to conduct meta-analyses such as that reported by Nissen and Wolski (2007), which combined data from multiple randomized clinical trials involving the same product. The results of this work will be used to inform FDA guidance as well as investigations that apply one or more Bayesian approaches to a single dataset and that compare the results with simulation to some of the newer methods using a frequentist paradigm.

Project team lead: Gary Rosner, ScD (Johns Hopkins)

For more information, please contact Gary Rosner at

Application of Common Data Elements To Improve Regulatory Science

This collaboration between the FDA and Johns Hopkins focused on the use of common data elements (CDEs) during the course of clinical trials. CDEs refer to a logical unit of data pertaining to one kind of information. They usually has a name, precise definition, and clearly enumerated values when applicable. CDEs help to ensure that data are captured and recorded uniformly and thereby both improve transparency and expedite the planning and initiating of clinical research. CDEs also assist comparison and aggregation of results across studies in systematic reviews and meta-analyses, thereby improving the efficiency of the scientific enterprise. This project systematically identified, developed, and disseminated CDEs with a focus on outcomes for use by sponsors and investigators engaged in eye and vision research. 

Through this process, the investigation generated fundamental new knowledge regarding the broader use of CDEs as they can be applied across other therapeutic areas and regulated products including prescription drugs (Center for Drug Evaluation and Research), biopharmaceuticals (Center for Biologics Evaluation and Research), and devices and radiologic products (Center for Devices and Radiological Health). As part of this initiative, researchers used literature reviews and consultations with glaucoma experts and patients to determine the most appropriate CDEs capturing relevant outcomes for research evaluating the efficacy and safety of minimally invasive glaucoma devices to facilitate bringing these innovative technologies to the marketplace. 

Project team lead: Tianjing Li, MD, PhD (Johns Hopkins)

For more information, please contact Tianjing Li at

Eliciting Patient Preferences To Enhance Regulatory Science

This collaboration focused on building capacity to measure and incorporate patient and caregiver preferences within the context of regulatory benefit-risk assessments. To make high quality decisions, the FDA not only requires information on the benefits and risks of therapies, but they need evidence on how to value this data. This said, evidence on how patients and caregivers value benefits and risk is often lacking. Increasingly, patient and caregiver values are being identified through the application of “stated-preference methods” to measure their priorities and preferences. These stated-preference methods can incorporate both qualitative and quantitative approaches; be used with large, diverse populations, including hard-to-reach patients and stakeholders; and their distributional characteristics can be assessed. This collaboration focused on advancing and disseminating methods for measuring the preferences of patients and caregivers. As part of this undertaking, Johns Hopkins and FDA staff and scientists collaborated on projects and initiatives to advance and apply stated-preference methods. 

Project team lead: John F P Bridges, PhD (Johns Hopkins)

For more information please contact John Bridges


The prevention of zoonotic diseases that have the potential to increase food animal mortality and emerge as human pathogens is of concern to FDA as well as other health and regulatory agencies in the US. Current policies are ineffective in preventing the dissemination of highly pathogenic avian influenza (HPAI) among domesticated avian species, as is clear in the current outbreaks in the US, since they have not prevented continuing large scale outbreaks of HPAI resulting in extensive economic losses, scarcities in eggs and egg products, loss of consumer confidence, and impacts on international trade. These policies are ineffective because they are based on an incomplete understanding of the challenges to biocontainment and biosecurity within this sector which includes layer and broiler chickens, ducks, and turkeys. There is an urgent need to incorporate state of the art knowledge of the industry in order to develop more effective policies and practices to prevent HPAI incursions, transmission, and potential emergence of strains that can infect humans. This project seeks to compile a review of the unrecognized gaps in biosecurity and biocontainment that pertain to industrial methods of poultry production (for both meat and eggs), to review the knowledge base documenting these gaps, and to evaluate existing and needed changes in management and operations within the industry.

Project team lead: Kathryn Dalton (Johns Hopkins)

For more information, please contact Kathryn Dalton at


Regulatory science should address the potential benefits as well as harms that matter to patients, but many clinical trials neglect harms. Even guidelines for assessing outcomes in specific conditions tend to focus on benefits rather than harms. This project compared methods for identifying potential harms and for selecting harms to include in research about patient preferences. This project aimed to help patients, consumers, and professionals make informed decisions about drugs by exploring different methods for identifying harms that are most important to patients.  It facilitated existing methods for social and behavioral research and it contribute to the development of larger studies (e.g., randomized controlled trials) to evaluate important methods for patient preference research. This project explored multiple sources of information that could be used to inform the selection of harms for measurement and to inform the weighting of benefits and harms in decision-making.  

Project team lead: Evan Mayo-Wilson, MPA, DPhil (Johns Hopkins)

For more information, please contact Evan Mayo-Wilson at


 A new class of commercial products that can deliver aerosolized nicotine entered the US marketplace in 2007. These products are called many things including: e-cigarettes, electronic cigarettes, electronic nicotine delivery systems, and vaporizers. E-cigarettes are a multi-billion dollar industry with millions of American users. This project will search patents to identify innovations, review patent content and translate findings and share with the FDA to support the FDA’s strategic plan including the need to ensure FDA readiness to evaluate innovative emerging technologies. 

This project sought to generate scientific evidence to inform e-cigarette product regulation, in anticipation of e-cigarettes being classified as tobacco products. Little is known about how product design may influence public health impact.  The research team used industry patents related to innovations around nicotine delivery.  The team quantified innovations related to e-liquids and nicotine delivery. Systematic searches identified relevant patent classification codes, which were then used to further narrow the scan.  Iterative searching focused on innovations related to nicotine and e-liquids.  Relevant patents were then reviewed using document analysis techniques to classify relevant industry strategies.

Project team lead: Ryan David Kennedy, PhD (Johns Hopkins)

For more information, please contact Ryan David Kennedy at