Skip to main content

Table 1 Comparison of randomized clinical trial and real-world data sources

From: Cardiovascular adverse events in oncology trials: understanding and appreciating the differences between clinical trial data and real-world reports

 Characteristic

Goal of Randomized Clinical Trial (RCT)

Real World Data (RWD)

(Possible confounders noted)

Treatment assignment

Randomization

Agreement to treat between patient and physician (Ideal might be limited due to insurance coverage or other considerations)

Concomitant medications

Specified in protocol

Some patients may be receiving concomitant medications excluded in the RCT protocol. Underreporting of concomitant medications more likely than in RCT

Risk/benefit

Data monitoring committees and cardiac panels scrutinize adverse events and can request further information on adverse-event experiences

Limited to the data in the database or individual event reporting. Insurance data may not include deaths as they are not a billable event

Adverse event reporting

Attempts are made to be complete

Events reported in structured format and actively collected as part of the ongoing clinical assessment

Reporting may be less rigid and may be over- or under-inclusive. Diagnosis often not verified, and a tendency not to report well established toxicities may also exist. History may be incomplete

Attribution of a cardiotoxic event to drug of interest

Not always possible to separate drug-induced change from a confounding factor

Not always possible to separate drug-induced change from a confounding factor. Inconsistent reporting of events may compromise the validity of conclusions

Medication and other compliance

Often recorded as per the protocol. Deviations in compliance exist.17

Data on drug compliance may not be sought

Clinical endpoints

Pre-specified efficacy and cardiac events adjudicated according to schedule in most trials

Adverse event noted clinically, and conditions of concern may be scrutinized. Endpoint data often not obtained

Confounders

Balance between measured and unmeasured confounders may be handled by randomization

Many unmeasured confounders allow for a more comprehensive overview of adverse events

Inclusion/Exclusion Criteria

Precisely defined in protocol

Patients may have comorbidities or take concomitant medications that are not allowed in the RCT and are not reported in database

Baseline variables

Attempts made to limit variables by inclusion and exclusion criteria

Often baseline variables not adequately measured

Cardiovascular data

Consistently measured using standardized techniques at defined intervals

Technique and timing may vary