Louise Parmenter, PhD

	Observational research studies have become an increasingly important source of evidence in the clinical research process. These studies provide valuable insight into the comparative effectiveness of a treatment in real clinical practice settings, and compliment data gathered in the trial environment, which can help key decision-makers to determine the quality and value of these products. Regulators, payers and providers have come to expect observational data to demonstrate evidence of value, and they will assess observational study quality when making decisions. The International Society for Pharmacoepidemiology (ISPE) has developed Guidelines for Good Pharmacoepidemiology Practices (GPP) in an effort to help sponsors develop the most effective studies to meet stakeholders’ needs.

	Yet, despite these guidelines and the growing demand for real-world evidence, many studies fail to meet required standards for design, operations and analysis, thus limiting their value to decision-makers. Perhaps part of the problem is that observational research requires clinical professionals to think out-of-the-box, and to be creative and innovative so that they can identify the right operational approach to gathering this data. This can be an unfamiliar place for researchers who are more accustomed to the rigors and structure of a clinical environment that the unpredictable nature of a real world setting.

	But if research teams want to meet the evidence expectations of key decision-makers, they need to move out of their comfort zone and seek out the best strategies to incorporate these studies into their broader research paradigm. That begins with making sure the research team has the expertise necessary to design and implement an effective study plan from the outset of the research program.

	Overall, there are three main challenges facing real-world evidence generation:

		Diversity of stakeholder needs within and across geographies

		Complexity of changing needs and varying real-world data access and availability

		Practical issues in gathering real-world evidence at the right time, at the right cost and to the right quality

	Without proper planning and a robust observational study strategy, these challenges can lead to fragmented and inefficient real-world evidence generation, with individual studies conducted by different teams to meet the requirements of individual healthcare decision-makers without taking into consideration the needs of stakeholders across the healthcare continuum.

	To build effective studies, planners first need to identify the needs of all key stakeholders, including regulatory bodies, treating physicians, payers and Health Technology Assessment (HTA) bodies. They can then marry these needs with insights into real-world data access and availability. This supports the design of a more streamlined program of real-world and late phase studies, that can generate more robust evidence in fewer studies for faster more cost-effective research.

	Organizations seeking to implement more global and impactful observational studies to support their clinical research initiatives should be open to new strategies and opportunities to engage patients in a real world setting.

	Using innovative approaches to design smarter studies focuses on four key elements:

		Observational study demand across the product life cycle, for example, with disease registries transitioning pre- and post-launch, or transitioning into a product registry, and an increase in requirements for post-authorization safety studies (PASS) and post-authorization efficacy studies (PAES) in Europe.

		Collaboration, with an increase in multi-sponsor registries, especially in orphan diseases where patient population is so limited.

		Patient centricity, with increasing inclusion of patient reported outcome (PRO) endpoints and ePRO technology, and direct-to-patient research.

		Greater healthcare data access and innovative study designs, including greater use of existing data (databases, claims data), and an increase in pragmatic trial designs and enriched studies that combine prospective and retrospective approaches in an efficient and effective manner.

	Remember, observational research isn’t just a nice-to-have addition to the clinical research program. It is an important source of evidence for healthcare decision-making, helping to reduce uncertainty by answering questions about treatment effects in real-world populations. Sponsors that plan for observational research plans from the outset of their studies, take advantage of experts in the field, and employ innovative study design methodologies will be able to gather the best data as efficiently as possible. That can save them time and money while generating a more robust set of evidence for the regulators, payers, providers and HTAs who will make critical decisions that directly affect the success of their products.
	
	

, Global Head of Operations, Epidemiology & Outcomes Research, Real-World & Late Phase Research, Quintiles
	

, Senior Director, EMEA Real-World & Late Phase Research Strategy Lead, Quintiles

Articles

Conducting observational studies is becoming a vital part of the clinical trial process, but research teams often struggle to operationalize these more creative and non-traditional study methodologies.

The Observation Study Conundrum


 Pharmacovigilance and safety of medicines are key priorities for regulatory agencies, with a strong focus on ensuring that both risks and benefits are monitored throughout a medicinal product’s lifecycle. For marketed products, sources of adverse reaction reports include those received spontaneously from health professionals as well as those derived from post-authorization studies and trials. Spontaneous reporting schemes are valuable tools for providing safety signals in a continuous manner, but they are known to be affected by many non-causal influences, including reporting stimulated by media events, blogging, etc. When signals arise, though, more formal research approaches are needed to provide a sound scientific context that can be used to confirm, refute, characterize and/or quantify possible safety concerns. Various study designs may be applied, ranging from active surveillance, observational pharmacoepidemiologic studies to clinical trials with a primary safety endpoint. Similarities and differences in research terms and approaches to post-marketing safety evaluation by the United States Food and Drug Administration (FDA) and European Medicines Agency (EMA) are highlighted here.


The FDA may require post-marketing studies or clinical trials to assess a known serious risk related to a drug, to assess signals of risk, or to identify an unexpected risk that available data indicates could be present. The April 2011 FDA Guidance for Industry on Post-marketing Studies and Clinical Trials

 gives FDA the authority “to require post-marketing studies or clinical trials at the time of approval or after approval if FDA becomes aware of new safety information. The guidance further states that “the authority to require a responsible person to conduct a post-approval study or studies or clinical trial(s) of the drug includes the authority for FDA to describe the study or trial to be conducted, including how the study or trial is to be done and the population and indication.” In other words, the FDA can require a study or clinical trial that is adequate to address the serious safety concern. Mandatory studies or trials are described as post-marketing requirements (PMRs). The term post-marketing commitment (PMC) is used to describe studies and clinical trials that applicants have agreed to conduct, but that are not a statutory requirement. Note the distinction between clinical trials and “studies” used by the FDA, which differs from the way the term “trials” is used in Europe.

Used to describe all required post-marketing studies or   clinical trials, including those required under FDAAA and those required   under subpart H of 21 CFR part 314, subpart E of 21 CFR part 601, the   Pediatric Research Equity Act, and the Animal Efficacy Rule

Used to describe studies and clinical trials that applicants have agreed   to conduct, but that will generally not be considered as meeting the   statutory purposes in 505(o)(3)(B) and so will not be required

Any prospective investigations in which the applicant or   investigator determines the method of assigning the drug product(s) or other   interventions to one or more human subjects

All other investigations, such as investigations with humans that are   not clinical trials as defined previously (e.g., observational epidemiologic   studies), animal studies, and laboratory experiments. Studies, in this sense, may be either   prospective or retrospective in design 

FDAAA states that, before requiring a trial, “FDA must find that a post-marketing study will not be sufficient to meet the purposes” of the request. This statement, together with the clear delineation between PMRs and PMCs, represents a formal change in FDA’s approach to post-marketing safety, with recognition that observational studies and routine data sources have a place alongside clinical trial data providing valuable evidence for post-marketing safety evaluation.    

PMRs under FDAAA generally would include, but not be limited to, the following: 

 Observational pharmacoepidemiologic studies

 Pharmacokinetic studies or clinical trials

 Drug interaction or bioavailability studies or clinical trials

Observational pharmacoepidemiologic studies are generally studies designed to characterize and/or quantify a serious risk associated with a drug exposure, or to evaluate factors that affect the risk of serious toxicity, such as drug dose, timing of exposure, or patient characteristics. To facilitate interpretation of the findings, the studies should always have a protocol, should include a comparison group, and should evaluate pre-specified research questions. However, comparison groups (often referred to as “control” groups) may be omitted when there is a scientifically valid reason to do so. For a solely descriptive study, the protocol may include clearly stated objectives for describing the safety issue, rather than research questions (e.g., estimating the upper bound for detectable risk). These study designs may be prospective, retrospective, or mixed. Data sources for observational studies include administrative healthcare claims data, electronic medical records, registries, prospectively collected observational data, or other sources of observational information. This gives a new, elevated stature to retrospective designs.

FDA provides examples of observational studies that include pharmacoepidemiologic studies designed to: 

 Estimate the risk of a serious adverse event or toxicity associated with use of a drug 

 Provide estimates of absolute risk (e.g., incidence rates) for a serious adverse event or toxicity

 Obtain long-term clinical outcome data, including information about potentially rare serious adverse events 

 Identify risk factors (e.g., patient characteristics) associated with the occurrence of adverse events among patients exposed to specified drugs

 Compare pregnancy incidence, pregnancy outcomes, and/or child outcomes after patient drug exposure compared to patients who did not receive the drug 

Meta-analyses may be designed to evaluate a safety endpoint by statistical analysis of data from more than one completed study or clinical trial. 

Clinical trials with a safety endpoint evaluated with pre-specified assessments and adequately powered to analyze the serious risk identified by FDA under section 505(o)(3) would be considered PMRs. Although efficacy and effectiveness endpoints may also be evaluated, the trial should be powered to adequately assess the safety concern that gives rise to the requirement. 

The inclusion of animal studies, which, in most cases, would have been completed pre-approval, recognizes situations where drugs may have received accelerated approval for serious or life-threatening conditions.   

In-vitro laboratory studies, pharmacokinetic studies or clinical trials, drug interaction, or bioavailability studies or clinical trials, might be used to address the needs of specific sub-populations, for example, patients with renal impairment. Additional uses include assessment of potential interactions of an approved drug with a frequently concomitantly prescribed medication. 

Table 2:  Examples of Post-marketing Requirements (PMRs) Under Section 505(o)

Observational pharmacoepidemiologic studies

Include registries (e.g., pregnancy, prospective   cohort studies, and retrospective case-control studies). 
 Assess a   serious risk associated with a drug exposure or to quantify risk or evaluate   factors that affect the risk of serious toxicity, such as drug dose, timing   of exposure, or patient characteristics.

Establish occurrence of all-cause mortality, cardiovascular death,   and cancer incidence and identify potential predictive factors in patients   treated with the drug compared to control therapies in all completed   randomized clinical trials that include the drug.
 Evaluate a safety endpoint by statistical analysis.

Examples are broad   and include asthma exacerbations, incidence of myocardial infarction in   patients treated with the approved drug in a follow-on trial after approval,   using the original randomized population, randomized withdrawal trials, Q-T clinical   trials, special population trials (pediatric, racial or ethnic group, vulnerable   population, pregnant women, patients with hepatic or renal impairment, HIV-1   patients), long-term cell and gene therapy product trials. 

Assess carcinogenic potential in appropriate species (e.g., mice and   rats) Assess the potential for reproductive toxicology in appropriate species   (e.g., monkeys or rabbits).

Include receptor affinity studies, drug resistance   studies, cross-contamination studies, immunogenicity assay studies. 

Pharmacokinetic   studies or clinical trials

Determine the optimal dose for maintenance therapy in patients with   chronic renal disease, a population at risk for drug accumulation.
 Study the pharmacokinetic profile in a rodent model of hepatic   dysfunction in order to evaluate the potential for toxicity in patients with   liver impairment.

Drug   Interaction or Bioavailability Studies or Clinical trials 

Assess in vitro whether products are p-glycoprotein   substrates and therefore could lead to increased drug concentrations and   toxicity.
 Assess potential interactions of an approved drug   with a frequently concomitantly prescribed medication.
 Evaluate whether multiple doses of an approved drug   alter the metabolism of a sensitive CYP2C9 substrate.
 Evaluate bioavailability of an oral drug in the   presence of food. 
 Evaluate   drug interactions or bioavailability when there are scientific data that   indicate the potential for a serious safety risk.

In the European Union, legal provisions covering pharmacovigilance and risk management fall under European Commission Directives. The main legal instruments are Directive 2001/20/EC2 and Directive 2004/27/EC3. The systems and controls described are also applicable to the member countries of the European Economic Area (i.e,. Iceland, Liechtenstein, and Norway). The legal text is supported by a series of guidelines. Of particular importance for post-marketing research is EUDRALEX Volume 9A4, a compilation of the major guidelines relating to post-marketing pharmacovigilance. 

In accordance with legal requirements, company sponsored post-authorization safety studies (PASS) may be required as a commitment at the time of market authorization or in confirming the safety profile of a medicine under conditions of normal use.  

From July 2012, the provisions relating to pharmacovigilance for marketed products will be further strengthened by revisions to the legal provisions. In addition to post-authorization safety studies (PASS), post-authorization efficacy studies (PAES) may be required by an authority at the time of granting the marketing authorization, or later.  

PASS include non-interventional studies performed to investigate a known or suspected safety issue as well as studies in which the number of patients to be included will add significantly to the existing safety data for the product.  Not to be confused with FDA definitions of “studies," PASS incorporate interventional clinical trials. In alignment with the United States, retrospective observational designs using claims databases and existing registries are among the types of PASS studies conducted in the EU.

Pharmacoepidemiological study or a clinical trial carried   out in accordance with the terms of market authorization, conducted with the   aim of quantifying a safety hazard relating to an authorized medicinal   product (Directive 2001/83/EC5 or 2004/27/EC 3)

Any investigation in human subjects intended to discover or verify the   clinical, pharmacological, and/or other pharmacodynamic effects of one or more   investigational medicinal product(s), and/or to identify any adverse reactions to one or   more investigational medicinal product(s) and/or to study absorption,   distribution, metabolism, and excretion of one or more investigational   medicinal product(s) with the object of ascertaining its (their) safety   and/or efficacy (Directive   2001/20/EC2).

A study where the medicinal product(s) is (are)   prescribed in the usual manner in accordance with the terms of the marketing authorization.   The assignment of the patient to a particular therapeutic strategy is not   decided in advance by a trial protocol but falls within current practice and   the prescription of the medicine is clearly separated from the decision to   include the patient in the study. No additional diagnostic or monitoring   procedures shall be applied to the patients and epidemiological methods shall be used for the analysis of   collected data (Directive   2001/20/EC2). 

Information on the design and conduct of PASS is provided in Volume 9A: Guidelines on Pharmacovigilance for Medicinal Products for Human Use, Sep-20084. This document provides a framework whereby a variety of data collection methods may be used to evaluate the safety of authorized medicinal products.  

Volume 9A describes a range of research methods under the headings:

Potential data sources are also described with recognition that both field studies and routine data sources may be used to retrieve the necessary data on exposures, outcomes, potential confounders, and other variables.

The importance of external validity is also mentioned. As far as possible, the study sample should be representative of the target population in which the safety concern has been raised. 

Table 4:  Epidemiological Methods for Post-Authorization Safety Studies (Summarized from Table 1.7.A, Volume 9A: Guidelines on Pharmacovigilance for Medicinal Products for Human Use, Sep-2008) 

Include sentinel sites, intensive monitoring   schemes, prescription event monitoring, and registries.  
 Disease/outcome registries may help collect data on   drug exposure and other factors associated with a clinical condition. A disease registry might also be used as   the basis for a case-control study.
 Exposure registries address populations exposed to   medicinal products of interest to determine impact.  These include pregnancy registries. Single   cohort studies measure incidence but have no comparator group and so cannot   provide proof of association. This   type of registry can be highly valuable when examining the safety of orphan   drugs. 

Comparative Observational Studies –   Cross-sectional

Cross-sectional studies (surveys) are limited in their application as   they are not able to address temporal relationship between exposure and   outcome. However, they can be useful   to examine disease prevalence and crude associations between exposure and   outcome in ecologic analysis.

Comparative Observational Studies – Cohort 

US and European Perspectives on Interventional and Observational Research Designs in Post-Marketing Safety

In the biopharmaceutical industry, research terminology largely reflects the phased regulatory drug development process. Around the time of product approval, there are many descriptors—late phase or late-phase research, registration/pre-registration research, peri/post approval research, postmarket research, Phase IIIb/IV research—but no formal definition exists, and the ones available are used inconsistently. 

These phase labels have served us well for many years, but they increasingly feel out of step with the new healthcare environment. Designed with the regulatory approval process in mind, the phase descriptions do not always fit with a broader stakeholder setting that includes physicians, payers, patients, and policy makers. For example, phase descriptions offer no insights into whether a product will be paid for by any healthcare payer.  

Late phase (with or without the hyphen) research is one of the most widely used terms, but it can be confused with late stage research which commences in Phase II of product development.  

Registration/pre-registration and peri-/post approval research are terms directly related to dossier submission and approval. In the pre-registration phase, all the necessary clinical trials have been completed and the drug is waiting for registration or approval for use by a governing body. Registered drugs are those registered or approved for use in a particular country, or group of countries. Approved drugs have received regulatory approval in an individual country or region, but are not yet on the market. Launched drugs are marketed in one or more countries. 

Postmarket research is a term favored by some divisions of the FDA, simply meaning in-label research on marketed products, while the term post-approval studies is used by the Centers for Devices and Radiological Health. 

One emerging construct is to use the term late phase to denote research on products following regulatory submission for at least one indication, and dividing this into Phase IIIb for trials occurring subsequent to that submission pre-approval or for new indications post-approval. Phase IV is utilized for studies conducted post approval that are not designed specifically to obtain a new labeling indication. 

Pre-approval Phase IIIb trials occur while the regulatory agency is reviewing the submission (Phase III pivotal data submitted) and usually focus on gathering additional data to support the product's anticipated launch, or commence long-term safety monitoring. Post-approval Phase IIIb trials investigate the use of an already approved product, outside the terms of their approved label, to expand the product indications. 

 Interventional designs test specific hypotheses (generally safety, effectiveness) relating to treatments for marketed products in existing indications. Observational designs examine actual use and often in more representative populations. Inclusion/exclusion criteria generally reflect an existing product label, although these may be based simply on product use in some observational studies.  

Bringing even more complexity to this confusing situation, many organizations are augmenting the drug development phase descriptions with new terms such as real-world, value demonstration, and comparative effectiveness research.  

There's no doubt that regulatory terms will continue to form the backbone of research terminology, yet adoption of new, consistent definitions are needed to help demystify late phase trials, and to facilitate a focus on research objectives to meet multiple healthcare stakeholder perspectives.  

Senior Director, Strategic Operations Quintiles Outcome E-mail: 

The adoption of new, consistent definitions are needed to help demystify late phase trials.

Louise Parmenter, PhD

Articles

The Observation Study Conundrum

US and European Perspectives on Interventional and Observational Research Designs in Post-Marketing Safety

Are Phase Labels Still Relevant?