Clinical trials remain the gold standard for interventional research in nursing, medicine, and other health professions. In the past, mobile technologies (MTs) have been used as interventions in clinical trials to treat individuals with diabetes, psychotic disorders, lung diseases, cardiovascular diseases, and other common problems (Triantafyllidis et al., 2019). Today, they are positioned to become important in clinical trials for collecting, tracking, and analyzing data (Perry et al., 2018). MTs can also be used to extract digital biomarkers (DBM) for the development of new, proxy end points in clinical trials or as predictors of progression of chronic diseases or development of new medical diseases (Clinical Trials Transformation Initiative [CTTI], 2016). In this column, I explore the selection and use of MTs as data collection tools capable of producing DBM for clinical trials.
MT is a broad term for applications, devices, wearables, ingestibles, implantables, and sensors to collect and track data of interest in any environment (Perry et al., 2018). The CTTI developed guidelines for the use of MT in clinical trials with a specific focus on meeting the requirements of the Federal Drug Administration. The guidelines cover all aspects of MT in clinical trials (see https://www.ctti-clinicaltrials.org/projects/mobile-technologies).
The aims and research questions of a clinical trial should drive the selection of MT (CTTI, 2018). In other words, MT must measure the outcome of interest. Although this statement seems obvious, researchers may be tempted to select technologies based on cost and convenience. A more thorough assessment of the specifications of MT is needed.
Candidate MT must have acceptable validity and reliability in order to be used in a clinical trial. The MT also must be acceptable to potential participants. Three aspects — validity, reliability, and acceptability — must be known before selecting an MT for use in a clinical trial, which means the research team must test the candidate MT before it is used. Resources are available to aid in the selection of MT, including the NOMAD project found at https://humanitarian-nomad.org/online-selection-tool and Kopernik Impact Tracker Tech at http://impacttrackertech.kopernik.info/tech-compare-features.
Researchers must also understand the raw and processed data of each candidate MT. For example, if a researcher wants to study the effects of stress on academic performance, a simple activity tracker can be used. These trackers collect raw data using a photoplethysmography, which sends flashing green LED lights onto the capillary bed of the wrist; the capillary bed absorbs more green light when the heart pumps blood but reflects green light back to the sensor during the resting phase (McGarry, 2018). These raw data are processed by the smartwatch as heart rate. However, if the activity tracker samples the heart rate every five minutes instead of every five seconds, the precision of the device may not serve the aims of the clinical trial. Furthermore, if the research team cannot import the raw and processed data into a database for additional analyses, then heart rate variability, which is needed to study stress, may not be available for the clinical trial.
MT must be integrated into a research data management system to ensure data integrity, enhance data security, and provide access to authorized research personnel (CTTI, 2018). The flow of data from the mobile device and/or its associated app to a server for storage, filtering, and processing must be planned to avoid errors in collection and transfer (CTTI, 2018). Researchers may need to consult with computer scientists, network specialists, or a technology company to make sound plans to support a clinical trial.
Data integrity is a critical concern in any clinical trial but even more so when participants are at a distance from the research team. To enhance data integrity with MT, the research team can require participant authentication when logging into the MT. Other methods to increase data integrity include having a timestamp on all collected data, encrypting data stored on the MT, and automatically backing up the data from the MT to a server.
Methods to monitor the quality of data for completeness, consistency, and correctness through automated processes must also be developed. If a smartwatch is used to measure the activity level of participants, an algorithm accounting for age, level of exertion, and percentage increase in heart rate above the resting rate could give researchers confidence in the quality of data if they fall into the expected range of values. On the other hand, when the data fall outside expected ranges or are missing, an automated notification system would reduce the burden of data monitoring. The research team should assign responsibility for investigating and taking action when a system notification is received.
The transfer of data from an MT to a server requires security measures to protect the data. The research team, in collaboration with a network specialist, can set up a secure network encryption certificate such as Secure Socket Layer and transmit data wirelessly using a File Transfer Protocol. Another method to secure data transfer is to require a matching pin on the server and device. When clinical trials are local, a mobile device can be plugged directly into the research server; once data are stored, they should be encrypted to protect all health and research information.
The research team needs to plan for methods to access the data stored on a server. The methods adopted in industry can be used in clinical trials. For example, requiring multiple layers of authentication to access data can reduce the likelihood of hackers gaining access to research data. Other measures to limit access include requiring a virtual private network to access the server, a password to access the data set, and an automated log of logins to the research server and data set.
Access to data given to members of the research team needs to be human readable, meaning the data should be summarized in a manner that researchers can readily understand. The CTTI (2018) recommends the use of a read-only dashboard to display meaningful data so no inadvertent access to raw data could compromise the integrity of the raw data.
DBMS FROM MTS
MT can produce data previously unavailable; therefore, researchers should consider how MT can produce DBM, defined as “objective, quantifiable physiological and behavioral data collected by digital devices” (Dorsey, 2020). Similar to traditional biomarkers, DBM can serve different functions in research: monitoring of conditions, predicting the development of conditions, assessing the risk of adverse health outcomes, diagnosing medical conditions, and monitoring the safety of and response to new medications (Babrak et al., 2019).
In a study of older adult mobility (Frith et al., 2019), a smartphone was used to collect data while participants performed the Timed Up and Go test and the 30-second Chair-Stand test. The traditional test would have produced only the time to complete the timed up and go and the number of stands from a seated position in a 30-second period. The smartphone app produced a vast array of new measurements. However, data produced from MT can be called DBM only if they are associated clinical or health outcomes.
The promise of DBM for use as clinical end points in trials is intriguing because researchers may be able to find biomarkers that can predict disease before it is manifest. However, the path to finding new DBM that can serve as clinical end points or proxy end points for clinical trials is long and must follow a rigorous process of validation. The CTTI (2018) group developed a publicly available resource detailing the steps for developing DBM as novel end points for clinical trials. Review the recommendations at https://www.ctti-clinicaltrials.org/files/detailedsteps.pdf and use the interactive tool at www.ctti-clinicaltrials.org/files/interactive-selection-tool.xlsx to analytically assess the value of different candidate DBM.
The use of MT in research is here. Researchers who use MT in their own clinical trials can reap the benefits of enrolling participants from different locations and from diverse backgrounds because MT can facilitate decentralized clinical trials. They can expect to benefit from efficient data collection in real-world settings, which is important for seeing the effect of interventions in the everyday lives of people. Finally, researchers can develop new DBM through rigorous testing so those biomarkers can be used in future clinical trials.
Babrak L. M., Menetski J., Rebhan M., Nisato G., Zinggeler M., Brasier N., Baerenfaller K., Brenzikofer T., Baltzer L., Vogler C., Gschwind L., Schneider C., Streiff F., Groenen P. M. A., & Miho E. (2019). Traditional and DBM: Two worlds apart? DBM
, 3(2), 92–102. https://doi.org/10.1159/000502000
Clinical Trials Transformation Initiative. (2016, May 22). Using mobile technologies to improve clinical trials
Clinical Trials Transformation Initiative. (2018). Steps for novel endpoint development, with suggested approaches and considerations
Frith K. H., Hunter A. N., Coffey S. S., & Khan Z. (2019). A longitudinal fall prevention study for older adults. Journal for Nurse Practitioners
, 15(4), 295–300.e1. https://doi.org/10.1016/j.nurpra.2018.10.012
McGarry C. (2018). Who has the most accurate heart rate monitor? Tom’s guide. https://www.tomsguide.com/us/heart-rate-monitor,review-2885.html
Perry B., Herrington W., Goldsack J. C., Grandinetti C. A., Vasisht K. P., Landray M. J., Bataille L., DiCicco R. A., Bradley C., Narayan A., Papadopoulos E. J., Sheth N., Skodacek K., Stem K., Strong T. V., Walton M. K., & Corneli A. (2018). Use of mobile devices to measure outcomes in clinical research, 2010–2016: A systematic literature review. Digital Biomarkers
, 2(1), 11–30. https://doi.org/10.1159/000486347
Triantafyllidis A., Kondylakis H., Votis K., Tzovaras D., Maglaveras N., & Rahimi K. (2019). Features, outcomes, and challenges in mobile health interventions for patients living with chronic diseases: A review of systematic reviews. International Journal of Medical Informatics
, 132, 103984. https://doi.org/10.1016/j.ijmedinf.2019.103984