Random controlled trials may be the gold standard in formal clinical research, especially for approval of medications and the development of clinical guidelines, but real- world studies are stealing the limelight as the practical tool for getting actionable answers on outcomes, safety, and many other health care endpoints.
“Random controlled trials are rigorous in terms of their methodology and that leads to results that have a lot of internal validity like a drug’s effectiveness in the study cohort, but those results may not present a complete picture of the drug’s outcomes in the general population,” says Mark Cziraky, vice president of HealthCore, a health outcomes company owned by WellPoint.
Real-world — also called observational––research has often been looked down on as a valid research methodology. Research snobs say that studies with uncontrolled variables measured by inconsistent methods yield unreliable results.
But others tout the contribution of observational studies, such as the famed Framingham Heart Study, which began in 1948 and continues today with the 3rd generation residents of that Massachusetts town. “That study advanced the techniques of risk assessment and also taught us the value of diet and exercise,” says Stephan Lanes, PhD, an epidemiologist at United Biosource, an observational research company owned by Medco.
“How do I bring real world evidence into the [drug] development process?” asks Jonathan Morris, MD, senior vice president at United Biosource, an observational research company.
Observational research plays a major role in comparative effectiveness studies and health plans rely on these studies to determine formulary placement for new medications or utilization management strategies for competing medicines.
HealthCore and United Biosource are involved in developments that are giving observational studies new roles in pharmacy management.
Beyond formulary management
Observational studies are expanding into drug development. “Much of our work is upstream from comparing medications; it’s in drug development,” says Cziraky. “We are using our capabilities to identify gaps in therapy and to quantify both the clinical and economic impact of those gaps.” The gaps he is talking about are where there is a need for a niche medicine.
“The clinical trials for FDA registration don’t address the commercial or utilization issues, and observational studies are filling that gap,” says Mark Cziraky, vice president of HealthCore, a health outcomes company.
”The focus today is on how to bring real- world evidence into the development process,” says Jonathan Morris, MD, a founder and senior vice president at United Biosource. “We caution drug companies that we don’t want to see regulatory successes and commercial failures. So, we study the populations a product will be used in or the comparators for a new medicine to aid in launching new agents.”
“The clinical trials for FDA registration don’t address the commercial or utilization issues and observational studies are filling that gap,” says Cziraky.
“For health plans our studies of new medicines help them interpret where the best place is for a new drug. Then we can monitor the drug to make sure that what they have projected or intended is actually how and where that medicine is being used,” says Cziraky.
Postmarket surveillance of drug and device safety in real-world conditions is another new area for observational research. The FDA’s Sentinel Initiative is working to create a monitoring system for safety problems. As currently envisioned, the Sentinel system will enable the FDA to capitalize on the capabilities of existing automated health care databases encompassing tens of millions of patient records.
Sentinel is an active surveillance system that uses periodic queries of frequently updated databases to target product safety information. Those databases will be tailored to include safety variables. Sentinel is viewed as a proactive safety initiative in comparison to passive surveillance through the current adverse event reporting system that is characterized by ?under-reporting and inconsistent information.
An old role for observational studies, outcomes research, is being revitalized. HealthCore has an agreement with the Delaware Medicaid program to evaluate the effects of benefit design, health policies, practice patterns, and geographic and demographic variations in utilization on outcomes. Its objectives are to provide insight to improve Delaware’s medical and drug policy decisions and develop new education and intervention programs.
The availability of larger and more powerful databases is a key reason observational studies are taking off.
There are widespread efforts by private companies and public organizations to develop massive research databases. WellPoint, HealthCore’s parent, and Medco, United Biosource’s parent, have huge databases that they can use in their studies. Both companies emphasize the use of firewalls to protect confidentiality.
One of the largest databases is managed by the Observational Medical Outcomes Partnership (OMOP), a public-private partnership chaired by the FDA and managed through the Foundation for the National Institutes of Health. OMOP conducts research across a network of 10 databases covering over 200 million patient lives to inform national drug safety efforts.
HealthCore and AstraZeneca have an alliance for building a database they call the Real-World Evidence Collaboration and the contract with Delaware Medicaid will provide additional records for that database.
AstraZeneca and IMS Health have announced a joint effort to develop an international database.
One of IMS Health’s strengths is in capturing physician-level prescribing data.
Observational databases are big in Europe, where they play a key role for Britain’s National Institute for Health and Clinical Excellence, NICE.
In addition to expanding horizontally, observational databases are getting deeper. “The research environment for observational studies is being enhanced by adding data from different sources like data from physicians’ offices, electronic health records, and in some cases chart reviews,” says Cziraky. “That allows us to ask more elegant and valuable questions for the end user. The goal is to get down to provider level or patient level data, as opposed to analyzing population level data.”
Data from multiple sources aids reliability. “As you add in data elements you can be more confident that what you think you are seeing is actually occurring,” says Cziraky.
For example, he says, health insurance claims are a major source of data for observational research and those records include diagnosis information. But those diagnoses are often aimed at providing billing information, not true clinical information. When a diagnosis from a second source is added you get a better picture.
All of this data expansion has another benefit. “Large databases are being segmented for different types of research such as looking at safety issues or clinical endpoints like hospitalizations or the incidence of acute myocardial infarctions,” says Morris.
While there are positive developments in observational research, at least one core issue remains. “Observational studies are gaining credibility, but the challenge is to strengthen the analysis of real world data to isolate variables or show how variables are related to results,” says Cziraky. “One step forward with prospective studies has been to add the strength of randomization and then step back to let the natural course of management take place.”
Indeed, random controlled trials are still the standard to get at cause and effect for one variable.