Patient reported outcomes (PROs) are increasing in use for both supportive and primary endpoint assessment in clinical trials. One of the most common ways to collect PROs is the use of a diary as well as other self-administered questionnaires. Historically, paper-based systems have been used, but these are often criticized for the lack of control in data entry when the study subject is not being directly observed. Detractors have voiced concerns regarding the lack of a documented audit trail when it comes to data entry and integrity; some reviewers have even stated or implied a preference for electronic PRO (ePRO) data capture systems.
As clinical trial technology continues to rapidly develop, advancing into the patient-facing arena, we can incorporate electronic diaries and other smart devices that collect real-time study subject responses with reliable user access and time stamps. This can go some way to placating the concerns of detractors who lean toward paper-based alternatives as subject responses are accompanied with time-stamps and electronic audit trails. Nevertheless, as we embrace technology we run the risk of disenfranchising some segments of the population.
Some ePRO vendors allow sponsors to purchase or rent electronic data collection devices and some even support B.Y.O.D. (bring your own device) platforms. These scenarios assume that study participants will either own or be able to use sophisticated electronic devices to record their assessments. Clinical trial ethics are based on medical ethics and one of the tenets of medical ethics is “Justice” – the belief that resources should be equally distributed amongst those who require them. If portions of the population are excluded from a clinical trial due the lack of access to a required device, the persons may be denied access to investigational trials. Therefore, the easiest way to increase access is to provide a loaner device with which the study participant could report outcomes.
Providing a technology-naive person with an unfamiliar loaner device is not the complete solution. An eNaïve study participant’s interface with an unfamiliar device may not be the same or similar to that of a more technologically savvy individual. Persons who are accustomed to technology may be less prone to use it with the same frequency and dexterity of someone more comfortablewith the medium. Even those with a relative technological comfort level may be influenced by the lack of familiarity with the device itself. Not only will the data collected from the two very different types of users differ, it will also be prone to unpredictable degrees of variability, with some study subjects potentially excluded from the study entirely. In an effort to avoid some of these biases, future stratification criteria may include technological dexterity and even android versus iOS.
Perhaps the best, albeit nonclinical, example of technology-based discrimination that resulted in selection bias and a skewed outcome is the 1948 US presidential election. A now infamous poll predicted that Thomas Dewey (Republican) would win the election and that the Democrat Harry Truman would lose. In the final tally the triumphant Truman had 114 more electoral votes and more than 2 million popular votes than the defeated Mr. Dewey.
The poll was designed to include quota sampling in an effort to avoid bias, but once the quotas had been met the remainder of those polled were interviewed by telephone. In 1948, more Dewey Republicans were likely to own telephones whereas Truman Democrats were more likely not to have telephone service. The poll predicted that Dewey would have 50 percent of the vote and Truman 45 percent; in the final count it was Truman who had the 50 percent and majority vote compared to Dewey’s 45 percent.
In designing our trials and selecting our outcome measures we must consider the population under study and the population of intended use. In rushing to embrace the reliability and robustness of ePRO collected data, we must be careful not to exclude important segments of society like the technologically naïve and economically disadvantaged. There are important benefits to ePRO systems but some patient populations who do not have experience and comfort with electronic devices might be better served with paper-based solutions. Although there is a growing preference for improved reliability of ePROs, there are situations where it may not be the best solution. Understanding the positive and possible negative aspects of data collection tools in special populations will help us to better select the most appropriate outcome measures and to defend those decisions.
*Thomas M Tremblay, RN BSN, is the Director of Clinical Development at Apexigen