Informed consent to participate remains a fundamental aspect of ethical clinical research. Potential participants of a clinical trial must be given adequate information about the study before they decide whether to participate in accordance with good clinical practice quality standards [1]. Providing informed consent means to agree to take part in the trial and to have understood what is involved, including the risks and benefits of participation [1]. Traditionally, the trial information is conveyed using printed documents that potential participants read before signing to indicate their consent to participate. The informed consent form (ICF), and the associated effective communication of study information, remains among the most challenging and complex processes within the clinical trial landscape. ICFs are known to have poor readability and take too long to be understood and digested effectively [2]. A review of ICFs developed for use in phase III oncology clinical trials showed that these were, on average, 21.4 pages long and that many participants had only a poor understanding of the key elements of their trial [2]. Poor understanding of the study requirements and treatment has been cited as a reason for early withdrawal from clinical trials [3]. To ensure that potential trial participants fully comprehend the study information, ICFs need to convey complicated and technical information in a way that meets the target group’s health literacy capabilities. ICFs have to maintain readers’ engagement sufficiently to ensure that they can make a fully informed decision on whether to participate.

In addition to patient-centered challenges of the ICF process, administrative aspects of the consenting process can pose challenges to investigators conducting clinical trials. Flawed informed consent processes are listed within the top 10 cited regulatory deficiencies and audit findings and are the third highest reason for US Food and Drug Administration (FDA) warning letters to clinical investigators [4-6]. Informed consent was among the top 2 most frequently observed issues in a recent auditing case study, conducted across 37 centers, and problems were identified related to processing errors and missing operational records [7]. Findings included missing signatures, incomplete ICFs, signing of incorrect ICF versions, and unauthorized site staff obtaining consents [7]. These are serious issues that can undermine the integrity of the consent process and the study, and they can result in the inability of researchers to analyze and report the data as intended.

Electronic consent (eConsent) uses digital technologies to enable the consenting process. Components can include multimedia to complement text-based content; interactivity (eg, to handle questions, test knowledge, explain definitions, and allow patients to resume the process from where they left off); electronic signature capture; status dashboards; and version control technology. eConsent aims to improve participant comprehension and engagement with study information and to address data quality concerns that may limit study integrity [8,9]. Although eConsent has been in use for about 15 years, its adoption has been slow until recently, when its accelerated uptake has been driven primarily by the COVID-19 pandemic [9]. Much of the supporting information on the promised benefits of eConsent comes from informal commentaries and reports from eConsent solution providers. In addition to digital technology, and just as with paper-based ICFs, eConsent solutions require good content to be effective. Similar to the computing analogy of “garbage in, garbage out,” poor eConsent content will result in poor overall effectiveness in terms of patient comprehension, acceptability, and usability, irrespective of the quality of the delivery technology.

The aim of our systematic review of peer-reviewed research was to provide a summary of qualitative and quantitative evidence to draw conclusions on the relative effectiveness of eConsent in comparison with traditional paper-based consenting.

The systematic literature review was conducted and reported in accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines [10]. A completed PRISMA checklist is included in Multimedia Appendix 1 [10]. We systematically searched the peer-reviewed literature for full papers and conference abstracts relevant to our review using Ovid Embase and Ovid MEDLINE on November 11, 2021. Ovid MEDLINE is equivalent in content to PubMed and additionally includes advanced search options (eg, adjacency operator and within-phrase wildcard) [11,12]. The search string contained terms related to electronic and consenting as follows: ([dynamic OR electronic OR interactive OR multimedia OR online OR tablet OR computer OR digital OR virtual] ADJ4 [consent* OR econsent OR e-consent]). Terms related to “electronic” were limited to the title, abstract, and keywords of a publication. The operator “ADJ4” was used to identify “electronic”- and “consent”-related terms separated by ≤3 words to filter for literature relevant to this review. The records were screened and selected based on our review of the title, abstract, and full text. No language restrictions or publication date limits were applied. The review was not registered, and a protocol was not prepared.

Publications reporting original, comparative data on the effectiveness of eConsent in terms of patient comprehension, acceptability, and usability were eligible for inclusion. Comparative data on the effect of eConsent on clinical study enrollment and retention rates, cycle time (ie, time taken to consent), site workload, and stakeholder views were also considered relevant. Head-to-head comparisons of paper-based methods versus eConsent were of particular relevance. Publications that did not present original data (eg, reviews, editorials, and commentaries) were excluded.

Following the systematic literature searches, duplicate records were removed using the deduplicate option in Ovid. All the remaining records were exported to EndNote X9 (Clarivate), and further duplicates were identified and removed manually. Two reviewers (EC and BB) independently assessed the systematic literature search results and the corresponding full texts following the initial screening of titles and abstracts, with one reviewer (AB) excluding the ineligible publications (eg, reviews), and the team of 3 reviewers resolved any disagreements by consensus-based discussions. Reasons for exclusion and inclusion were captured.

Data extraction (conducted by AB and reviewed by BB) included measures and outcomes for patient comprehension, acceptability, usability, enrollment rates, retention rates, cycle time, site workload, and stakeholder views. The extracted data were summarized descriptively. Data on patient comprehension, acceptability, and usability with eConsent versus paper-based ICFs were tabulated as part of the main descriptive summary. An overview of all the studies identified for inclusion is provided in the Multimedia Appendix 2 [13-49].

For studies comparing patient comprehension, acceptability, and usability with eConsent versus paper-based ICFs, we estimated the quality of the evidence by categorizing their methodological validity as “high,” “moderate,” or “limited.” The study methodologies that we categorized as having high validity (score=+++) were those that used comprehensive assessments including detailed and open-ended questions (eg, “Tell me what will be done during the study visits”), possibly using established instruments as part of the formal assessments. Methodologies that involved self-rating by participants (eg, “Did you understand the following aspects of the study?”), without formal testing, were categorized as having moderate validity (score=++). When a methodology that involved limited questioning was used in the studies or when methodological details were not reported, we categorized these studies as having limited validity (score=+).

The systematic literature search identified 1872 publications (Figure 1). Of these 1872 publications, 608 (32.48%) duplicates were excluded before screening, and a further 1228 (65.6%) were excluded based on screening by title, abstract, and full publication, with the most common reason for exclusion being that the publication did not report on eConsent research. A total of 36 studies met the eligibility criteria [13-48], and an additional outcomes publication [49] was retrieved manually based on the identification of its accompanying methodology article during screening. Thus, in total, 37 publications (32 full publications and 5 conference abstracts) were included in this review (Multimedia Appendix 2) [13-49].

The included publications together described 35 studies (2 studies were each covered by 2 publications). Most of the studies (28/35, 80%) were from North America (United States: n=26, 93%; Canada, n=2, 7%; Multimedia Appendix 2). The remaining studies (7/35, 20%) were from Europe (Italy: n=1, 3%; Ireland and United Kingdom: n=1, 3%; United Kingdom: n=1, 3%), Australia (n=2, 6%), and Gambia (n=1, 3%), and 1 (3%) was multinational. Taken together, these studies included a total of 13,281 participants. The number of participants per study ranged from 9 to 3485. In total, 13 (37%) out of 35 studies were conducted as part of randomized (n=10) or nonrandomized (n=3) clinical research studies, 14 (40%) studies were simulated consent studies, and 8 (23%) studies were survey or interview studies. Most of the research and simulation consent studies (23/27, 85%) were conducted in person (Multimedia Appendix 2). Comparative data on patient comprehension, acceptability, and usability with eConsenting were provided in 26, 13, and 6 studies, respectively, of which 20, 8, and 5 studies included comparisons for eConsent versus paper-based ICFs, respectively. Aspects of eConsent in relation to enrollment rates, retention, cycle time, staff workload, and stakeholder views were covered in 12, 1, 13, 3, and 5 studies, respectively. Age groups ranged from 8 years to 91 years in the 14 studies that included age range information. Among the 23 studies that provided sufficient information on average (mean or median) age, the average age was <50 years in 12 studies and ≥50 years in 11 studies.

Overall, 26 studies (8778 participants in total) assessed the aspects of patient comprehension with eConsenting.

Overall, 13 studies (1694 participants in total) assessed the aspects of patient acceptability with eConsenting.

Overall, 6 studies (582 participants in total) assessed the aspects of patient usability with eConsenting.

A total of 12 studies (6399 participants in total) assessed the effect of the ICF format on the aspects of patient enrollment, with mixed results. Comparisons of consenting rates with eConsenting versus paper-based ICFs were reported in 5 studies [15,18,28,35,41]. In the study by Bickmore et al [15], a significantly higher proportion of participants in the eConsent group than those in the paper group signed their ICFs (P=.01). Consenting rates were also higher with eConsent than with the paper-based ICFs in a study by Chalil Madathil et al [18] (P value not reported). Consenting rates were similar between the groups in a study by Jeste et al [28] (P value not reported), and Rothwell et al [35] reported that consenting rates were similar in paper-based ICFs and video eConsent groups, but the rates were lower in the app eConsent group (P value not reported). In a study by Simon et al [41], enrollment was significantly higher with a face-to-face informed consent process than with eConsenting (P=.004), although immediately after the consenting process, similar proportions of the 2 groups had reported their intention to enroll; the eConsent process was conducted at the same location as the face-to-face process.

Overall, 4 studies reported eConsenting rates using different electronic media formats [21,23,24,33,38]. No significant differences in enrollment rates were observed with animated versus text-based video consents by Naeim et al [33], with video- versus text-based consenting by Fanaroff et al [21], or with different levels of eConsent interactivity by Golembiewski et al [23] and Harle et al [24]. Siegel et al [38] observed an increase in enrollment rates after a content redesign and attributed the increased rates to the web-based consenting being directly integrated with new patient on-boarding.

Furthermore, 3 studies described results about preferences and found that the format of the ICF made little difference to participants’ decision-making regarding study participation [13,17,49]. In the study by Abujarad et al [13], participants were asked to score the importance of the consenting process in their decision to participate; scores were not significantly different between the paper ICF and the eConsent groups. In the study by Knapp et al [49], similar proportions of patients in the paper-based ICF and the eConsent groups found that the trial information provided helped them make their decision about whether to take part [49]. Study researchers who were surveyed in the study by Cagnazzo et al [17] thought that the use of eConsent had little influence on whether patients declined to participate in a stud.

None of the included studies reported overall study retention comparisons between the 2 consenting approaches of paper-based ICFs versus eConsenting. Fanaroff et al [21] (3485 participants) assessed different formats of eConsenting and found no statistically significant differences in the proportions of enrolled patients who subsequently completed the 2 requested study procedures, namely, a blood draw and survey questions.

A total of 13 studies (2063 participants in total) assessed cycle time, and 10 studies (covered in 11 publications) assessed the comparative effect of eConsent versus paper-based ICFs on consenting times, 2 studies (3 publications) asked about perceived consenting time, and 1 study assessed consenting times with different electronic formats. eConsenting took more time than paper consenting in the studies by Chapman et al [19,20] (P=.006), Jayasinghe et al [27] (P<.001), McCarty et al [30] (P<.001), Rowbotham et al [37] (P<.001), Simon et al [39] (P<.001), Sonne et al [42] (P value not reported), and Varnhagen et al [45] (P<.001; partial η²=0.36). eConsenting was faster than paper consenting in the studies by Afolabi et al [14] and Jeste et al [28] (P value not reported in either study). Chalil Madathil et al [18] found no significant effect of consenting condition on time taken to complete the task. Abujarad et al [13] and Warriner et al [47] asked participants about their perceived time to complete the task and found no statistically significant differences between the eConsent and paper-based ICF groups. Different electronic formats of eConsenting did not significantly affect consenting times in the study by Golembiewski et al [23] and Harle et al [24].

In total, 3 studies (3284 participants in total) assessed the site workload. Hospital staff in the study by Chalil Madathil et al [18] reported a less subjective workload with eConsenting than with paper-based formats (P=.02), and the responses were assessed using the National Aeronautics and Space Administration Task Load Index. Site advisory group feedback in the study by Vanaken and Masand [44] included a beneficial reduction in the administrative burden and reduction in paper trail, although a potential for increased workload was also noted, for example, in relation to training and device management. In the study by Zeps et al [48], clinical trial researchers noted that eConsent devices could be clunky and prone to malfunction, which increased overall study time and burdened trial staff.

Overall, 5 studies (3416 participants in total) assessed stakeholder views. Staff in the study by Chalil Madathil et al [18] preferred eConsenting formats over paper-based consenting (differences among systems, P<.005). In the study by Warriner et al [47], the findings from a telephone survey of practice sites that administered both consent processes favored eConsent over paper-based ICFs, but the differences were not statistically significant. Health Authority representatives were in favor of the broad implementation of eConsent in alignment with local regulations in the study by Vanaken and Masand [44]. However, approximately half (53%) of the surveyed research participants preferred having both a paper document and an eConsenting system [44]. Similarly, most centers (65%) in a survey by Cagnazzo et al [17] preferred using a paper-based ICF in parallel with eConsenting. Clinical research stakeholders surveyed by Cagnazzo et al [17] in late 2020 thought that at a regulatory level, the use of eConsent might increase the time to study approval. In the survey of clinical trial researchers’ opinions on eConsent conducted by Zeps et al [48] in early 2019, a total of 68% of the respondents believed that ethics committees would not approve the use of eConsent or were unsure if they would, while 67% of the respondents thought that the lack of standardized, consistent guidance across the sector was an important barrier to success and 60% of the respondents believed that the high initial cost might be a barrier to uptake.

Our systematic literature review aimed to assess the effectiveness of eConsent in terms of patient comprehension, acceptability, usability, study enrollment and retention rates, cycle time, and site workload, primarily in comparison with traditional paper-based consenting. We identified 37 primary publications for inclusion that together described 35 studies (13,281 participants in total). Our results showed that compared with patients who used paper-based consenting, patients who used eConsent had a better understanding of the trial information, showed greater engagement with content, and rated the consenting process as more acceptable and usable. Cycle times were increased with eConsent, potentially reflecting the greater patient engagement with the content. Data on enrollment, retention, and site workload effects were limited. Some general themes emerged in relation to the effectiveness of eConsent, its administrative aspects, and the variability in eConsenting formats used across studies. We have discussed these under the following subheadings.

The types of eConsent used varied considerably across the included studies. Formats included straightforward digitization of paper documents, signature management systems, audio- and video-enhanced content, and fully interactive systems. Across formats, active multimedia engagement principles were observed, and significantly improved comprehension was achieved with highly interactive versions compared with less-interactive eConsent versions [22]. Animated video–based information was found to be easier to understand than text-based videos [33]. Even for text-based formats, use of line spacing, bold font, and bullet points could improve comprehension [34]. This variability in format has implications for future work, which might explore the differences and most effective formats further.

There was also variability in terms of how comprehension, acceptability, and usability were assessed. Among the studies that assessed the effectiveness of eConsenting compared with paper-based consenting, half of the 20 studies on comprehension had high methodological validity, but only 1 of the 8 studies on acceptability and 1 of the 5 studies on usability did so.

In its guidance on the use of electronic informed consent, the FDA notes the following:

In line with the FDA guidance, we suggest that a consenting format should be referred to as eConsent only if it can support patient engagement using multimedia components (eg, text, graphics, audio, and video) together with interactive functionalities to share information related to the study. If a digital consent solution does not have these capabilities, we suggest that it should be referred to as a digital consent form rather than a true eConsent solution.

Limitations of our systematic literature search include the fact that the search strategy that we used would have missed some potentially relevant studies while trying to keep the number of publications for screening manageable. Among eConsent studies, it is conceivable that there may be a reporting bias in favor of those finding comparative differences. Differences in the outcome measures used, including differences in their validity, made comparisons between studies challenging and pooling across studies unfeasible. Most studies included in this review did not provide a detailed description of the eConsent format, and such information should be included in future studies to allow researchers to assess and compare the results across studies. These observations are a call to action to harmonize the analysis, documentation, and reporting of eConsent findings, as well as the parameters defining best practices for eConsent (including whether these are met by current eConsent vendors). The same applies for the used terminologies and processes around eConsent. A current ongoing initiative of the European Forum of Good Clinical Practice aims to achieve the standardization within the clinical trial [69].

In conclusion, this systematic review showed overall patient benefits with eConsent versus paper-based consenting in terms of understanding, acceptability, and usability. No study reported significantly better overall patient benefits with paper-based ICFs than with eConsent. eConsenting can increase enrollment into clinical studies by improving access to research. Comparative data from site staff and other study researchers indicate the potential for reduced workload and lower administrative burden with eConsent. In addition to these benefits, there are various other advantages associated with the use of digital solutions, including preventing flawed consenting processes, ensuring data quality, and supporting study integrity. Importantly, there are several avenues for future research that we believe are necessary. These include, but are not limited to, research that explores the best methodologies to target specific measures of eConsent efficacy (eg, recruitment, retention, and site experience); research that explores cross-cultural and globalization elements of eConsent; and research that makes better use of the theoretical underpinnings for why eConsent methods are more efficacious.