Real-World Data Should Not Replace the Need for Clinical Trials

^{Photo/Thumb have been modified. Courtesy of spainter_vfx - stock.adobe.com.}

Using real-world evidence presents opportunities to supplement or replace some clinical trials. But recent research revealed that only 15% of clinical trials published in 2017 could feasibly be replicated using currently available real-world data sources.

Despite the increasing use of real-world evidence in medical research, observational methods are not likely to replace the need for traditional clinical trials, according to the findings of a study published in JAMA Network Open.

“Using observational methods and data sources is not feasible for examining the safety and efficacy of a medical product that is not in widespread use in routine clinical practice,” the researchers wrote.

The objective of the study was to identify the number of high-impact journals that could be feasibly replicated using observational data from insurance claims and electronic health records (EHRs). Researchers measured the number and percentage of trials for which the intervention, indication, trial inclusion and exclusion criteria and primary end points could be found from those records.

The research team identified the top seven high-impact journals in the category of medicine, general and internal, defined as journals with an impact factor greater than 10 in 2016. The seven journals included: New England Journal of Medicine, Lancet, JAMA, The BMJ, Annals of Internal Medicine, JAMA Internal Medicine and PLoS Medicine.

To start, the researchers examined the study intervention, followed by the indication or condition being treated, then the trial inclusion and exclusion criteria identified from the trail registration on ClinicalTrials.gov. Then, the research team examined the trial’s primary end points.

Using descriptive statistics, the researchers characterized the percentage of trials where the aforementioned data were included in the insurance claims or EHR data.

In total, 429 clinical trials were published in high-impact journals, of which, 220 made up the final sample.

Of the 220 trials, 86 (39.1%) had an intervention that could be routinely ascertained from observational data. Common interventions that could not be ascertained from EHR or claims data included new drugs that had not yet secured U.S. Food and Drug Administration (FDA) approval (34.1%), educational, behavioral or procedural interventions (14.1%) and medical device trials (6.4%).

Replication of trials of FDA-regulated pharmaceutical interventions were often precluded by the lack of approval at the time of the trial publication, the study authors wrote.

Out of the 86 trials where an intervention could be discovered, 62 (72.1%) had indications that could also be routinely ascertained from EHR or claims data.

In all, 33 trials generated evidence that could be routinely ascertained from observational data.

The findings reinforce the caution about the promise of real-world evidence replacing randomized clinical trials, the authors added.

Observational data can complement, but not completely replace the trials, the authors concluded.

“Further development of observational methods and data systems may help realize the potential of (real-world evidence) and may, in turn, translate into more generalizable medical research and more rapid evaluation of medical products,” the authors wrote.

Get the best insights in digital health directly to your inbox.

Related

Real-World Data, EHRs Can Analyze and Compare Treatment Practices

38M Patients’ EHR Data Replicated Clinical Trial Results

Smart Shirt, Real-World Data Shine Light on Surgeon Stress