The Effect of Digital Literacy on Participation in Social Media Clinical Trials in Cancer: Tailoring an Informed Consent Process

Abstract

Introduction:

This study asked: (1) How does digital literacy influence one's decision to consent to a social media intervention study? (2) What is a brief way to assess individual digital literacy before an individual's decision to participate in a trial? and (3) How can a consent process be tailored around an individual's digital literacy level?

Methods:

We used an assessment tool to investigate digital literacy of those who chose to consent to a clinical trial and those who did not consent to the clinical trial but agreed to participate in a digital literacy study.

Results

A total of 161 hospice caregivers completed the digital literacy assessment. Older individuals and those who rated themselves as more proficient in the use of technology and social media were more likely to consent to the social media clinical trial.

Conclusions:

We found that asking participants to rate their technology skills and social media skills allows researchers to tailor a consent process. For those who are comfortable with technology and social media the traditional process is appropriate. For individuals that rate themselves with weaker technology and social media skills it is important that the consent process includes assurance they will receive adequate support in the use of the technology and the media. The next step is to test the assessment and tailoring of consent processes for a social media clinical trial. Clinical Trial # NCT02929108.

Introduction

Cancer research raises many ethical issues, especially related to recruiting and retaining vulnerable human subjects at the end of life. Ethical issues in this population include challenges in recruitment protocols, informed consent processes, the appropriateness of some study designs, and potentially burdening a seriously ill population.^1–8 Although ethical guidelines for research in hospice and palliative care have been recommended, such recommendations do not reflect consensus among all stakeholders.^9,10 Published recommendations emphasize the need for open inclusion criteria, education for providers and others on the value of research for this population, strategies to decrease gatekeeping by professionals, and recruitment and retention approaches that feature an opt-out rather than an opt-in strategy.^11–13

Family caregivers of hospice cancer patients are particularly vulnerable; our research has shown that these family caregivers have a poorer quality of life and are more depressed and anxious than those of noncancer hospice patients.^14,15 A narrative review of 46 articles on ethics in hospice and palliative care research identified numerous ethical considerations, including the importance of respecting patient and family autonomy and their desire (or not) to participate in research.¹⁰

It is noteworthy that, although challenges exist in recruiting for hospice research, 46% of hospice patients and 60% of their family caregivers value the opportunity to participate in research and contribute to improving future care.¹⁶ Many cancer research participants participate in research, so others may potentially benefit.¹⁷ Remarkably little has been written about the ethical conduct of hospice clinical trials, and very few studies have examined the ethical issues introduced by social media interventions in hospice clinical trials.

Social media is defined as electronic communication that creates online communities to share information.¹⁸ Social media platforms have gained worldwide acceptance, linking strangers together in new ways. Social media can effectively disseminate cancer-related information to family caregivers and create a social network of individuals facing similar situations.^19,20 Family caregivers of hospice cancer patients become especially isolated as their social support networks shrink over the course of hospice enrollment.²¹

Social media platforms offer these isolated and vulnerable caregivers a way to share their experiences with others in similar circumstances. In contrast, the introduction of such new tools may bring challenges for those who do not have previous experience with or feel comfortable in the use of information and communication technologies (ICTs). The ability to effectively use ICTs is often referred to as digital literacy.²²

Although new digital health technologies emerge for cancer patients to access information, communicate with peers and clinicians, and track their symptoms, patients with limited digital literacy are often excluded from these innovations and existing disparities in cancer care may be further exacerbated.²³ Those who do not understand and do not have the skills to use technology-based health care interventions, despite its availability, face a new kind of disparity and bias.

Digital literacy can be a barrier to care in the same way any traditional disparity or health literacy deficit can be. Although the digital divide on technological access is closing, the need for sensitivity around literacy on the use of those resources remains a consideration with many in the population. Thus, the assessment of an individual's digital literacy and the making of accommodations related to any resulting barriers is a new challenge that must be addressed as with the development and testing of any new health technology intervention.

ACCESS (ACCESS FOR CANCER CAREGIVERS TO EDUCATION AND SUPPORT FOR SHARED DECISION-MAKING)

This article reports on a supplement supported substudy of a larger National Cancer Institute-funded clinical trial called ACCESS (R01CA203999). ACCESS is a 5-year cluster-crossover pragmatic randomized clinical trial testing an intervention that provides education and emotional support to family caregivers of hospice cancer patients based on a structured process for shared decision-making.¹⁴

ACCESS has three components: (1) Facebook groups to provide education and emotional support to family caregivers (offered to Group 1 and Group 2), (2) web-conferencing to involve the caregivers in hospice care plan meetings (offered only to Group 3), and (3) a structured shared decision-making process to guide team discussion (also offered only to Group 3). These components are designed to facilitate shared decision-making between hospice staff and family caregivers, lower caregiver anxiety, and increase caregiver knowledge regarding pain. Seven hospice agencies have been randomly assigned to one of three groups in the cluster-crossover clinical trial.

Participants enrolled in hospices assigned to Group 1 receive usual hospice care and thus does not experience any of the three intervention components. Participants enrolled in hospices assigned to Group 2 receive usual care and the first intervention component by participation in a facilitated hidden (private) Facebook groups for informational and emotional support. Participants enrolled in hospices assigned to Group 3 receive usual care and all three intervention components as they participate in Facebook groups, and attend (through web conferencing or telephone) biweekly hospice interdisciplinary team meetings where they take part in shared decision-making.

Each agency is assigned to each condition for 12 months with a 90-day washout period between each crossover to prevent contamination between conditions. Four hundred ninety family caregivers of hospice cancer patients enrolled in the study. Of those who declined to participate, nearly 25% stated a reason related to technology as the reason.

BENEFITS VERSUS BURDENS THEORETICAL FRAMEWORK

The study presented here is built upon a theoretical framework developed by Ulrich et al. noting both the benefits and burdens of research participation in cancer clinical trials.¹⁷ The framework holds that a decision to participate in research involves the participant's perceptions of benefits and burdens that are weighed by participants before decision-making. The decision is based upon the potential participant's perceptions of benefits and burdens related to factors in five dimensions: physical, psychological, economic, familial, and social.

In this study, Physical factors relate to the (broadly defined) physical requirements of study participation, such as the use of Facebook and the technology around it. Psychological factors involve the potential participant's positive and negative attitudes or feelings related to participating in research. Economic factors pertain to the financial cost (actual or perceived) required for participation, including the cost of the required technology. Familial factors take into consideration the effect(s) of research on the family. Finally, social factors are tied to how individuals perceive their social networks benefit or are burdened by the research.

This project seeks to examine the impact of digital literacy on consent into ACCESS. The substudy asks the following research questions: (1) How does digital literacy influence one's decision to consent to a social media intervention study? (2) What is a brief way to assess individual digital literacy before an individual's decision to participate in a trial? and (3) How can a consent process be tailored around an individual's digital literacy level? We hypothesized that those who consented to the trial would have a higher digital literacy than those who did not consent.

Methods

We assessed individual digital literacy for both those who consented to the ACCESS clinical trial and for those who did not consent to ACCESS but consented to this substudy. All participants were caregivers of hospice cancer patients in one of seven hospice agencies in one Midwestern state. Using an Institutional Review Board (IRB)-approved protocol, research staff were given contact information for caregivers of hospice patients enrolled in a participating agency. They contacted the caregivers and explained the ACCESS clinical trial and asked for consent. If the caregiver consented to participate in the clinical trial, they were sent a link to a REDcap survey that included a digital literacy measure along with other measures taken at baseline for the larger trial.

If a caregiver did not consent to participate in the clinical trial, they were asked if they would consent to participate in a short survey about their use of social media. If they agreed, they were sent a REDcap link to a survey focused solely on their digital literacy. Informed by a recent review of digital literacy assessment instruments,²⁴ we modified and combined two published scales for our study (Table 1). The established scales, already psychometrically tested for reliability and validity,^25,26 are intended for audiences across the lifespan.

Table 1.

Brief Digital Literacy Assessment

Demographic Questions:

What is your age? What is your gender? What is your race?

Self-Assessment of Skill

Q1. On a scale of 1–10 with 1 being poor and 10 being excellent, how do you rate your overall skills with technological devices such as cell phones and computers?

Q2. On a scale of 1–10 with 1 being poor and 10 being excellent, how do you rate your overall skills with social media applications such as Facebook or Twitter?

Experience with Social Media

Q3. On which of the following social media do you have an account? (note all that apply)

None (if none go to question#13) Snapchat

Facebook Twitter

Instagram Other

Q4. Which of the above do you use most often?

Social Media Usage

Q5. Do you check your account at least daily? Yes No

Q6. Do you check your account from a cell phone? Yes No

Q7. Do you post status updates to your account? Yes No

Q8. Do you post photos to your account? Yes No

Q9. Do you read others' postings? Yes No

Q10. Do you browse posts and photos of others? Yes No

Q11. Do you comment on others' posts, updates, or photos? Yes No

Q12. Do you click “like” or react to others' posts, updates, or photos? Yes No

Benefits, Barriers, and Privacy

Q13: What, if anything, do you enjoy or think is good about using social media?

Q14: What, if anything, do you dislike or think is bad about using social media?

Q15: What are your concerns about privacy when using social media?

DATA ANALYSIS

Before testing our hypotheses, we examined measures of central tendency (i.e., means, standard deviations, and frequencies). To test our hypotheses regarding digital literacy and enrollment, we used Wilcoxon ranked sum tests due to assumption violations and, for digital literacy and age, we used Kendall's tau correlation tests due to assumption violations and ties within the data. To determine magnitude of the correlations, we examined the means of each group (enrolled and not enrolled) for, the Wilcoxon tests and for Kendal's tau, we examined the correlation coefficient. We considered p-values <0.05 to constitute statistical evidence of a relationship between the two variables in their respective tests.

The responses to the open-ended questions were downloaded for categorization and review by the research team. Similar responses for each question were grouped together and tallied for descriptive purposes.

Results

The sample for this study consisted of a subgroup of consented (n = 111) and nonconsented (n = 50) participants to ACCESS, resulting in a total sample size of 161. They were predominantly female (77%) and white (74%) with a mean age of 56.6 years. A descriptive summary of the sample is provided in Table 2.

Table 2.

Summary of Digital Literacy, Enrollment, and Participant Age

	OVERALL (N = 161)
Technological skill
Mean (SD)	7.20 (2.36)
Median [min, max]	8.00 [1.00, 10.0]
Social media skill
Mean (SD)	6.03 (3.08)
Median [min, max]	7.00 [1.00, 10.0]
Usage
Mean (SD)	6.19 (2.00)
Median [min, max]	7.00 [0, 8.00]
Social media platforms
Mean (SD)	1.63 (1.28)
Median [min, max]	1.00 [0, 5.00]
Enrolled
No	50 (31%)
Yes	111 (69%)
Age (years)
Mean (SD)	56.6 (13.5)
Median [min, max]	58.0 [19.0, 81.0]
Gender
Male	37 (23%)
Female	124 (77%)
Race
Black/African American	36 (22%)
Other	6 (4%)
White/Caucasian	118 (74%)

The Wilcoxon tests for digital literacy and enrollment indicated that for all aspects of digital literacy except usage, there were mean differences between enrolled and not enrolled participants. For technological skill, on average, enrolled participants rated themselves a mean of 7.86 out of 10, whereas nonenrolled participants rated themselves 5.74 out of 10 (p < 0.001). For social media skill, on average, enrolled participants rated themselves 6.99 out of 10, and nonenrolled participants rated themselves 3.9 out of 10 (p < 0.001).

In addition, on average, enrolled participants used a mean of 1.85 social media platforms, whereas nonenrolled participants used a mean of 1.12 (p < 0.01). Figures 1–3 illustrate the distribution of technical and social media skill by group as well as the number of social media platforms. The Kendall's tau correlation tests for digital literacy and age and indicated that age was significantly correlated with all aspects of digital literacy (Table 3), as expected older persons have lower digital literacy scores. Specifically, older age resulted in lower self-rated technological skill, self-rated social media skill, social media usage, and number of social media platforms used.

FIG. 1.

Consented and self-rated technological skill.

FIG. 2.

Consented and self-rated social media skill.

FIG. 3.

Consented and platforms used.

Table 3.

Participant Age and Digital Literacy

DIGITAL LITERACY	AGE (τ)	p
Self-rated technological skill	−0.35	<0.001
Self-rated social media skill	−0.32	<0.001
Social media usage	−0.24	<0.001
Social media platforms	−0.36	<0.001

The open-ended questions revealed the primary reason individuals such as social media, their dislikes of social media, and their privacy concerns related to social media. There were 131 (81%) of the respondents who provided a reason for liking social media. Of those responding, the most common reasons people liked social media was to stay connected with friends and family and share photos (n = 96; 73%), and the second most noted reason was to get information (n = 12; 9%). There were no patterns identified for liking social media in the remaining 23 (18%) responses.

There was a larger variance in the noted dislikes of social media, which was noted by 116 (72%) of the sample. The largest number of dislikes was in the other category (n = 52; 44%). These responses were so dissimilar they could not be categorized. The most common dislike was related to the amount of personal information many individuals choose to share, as 22% (n = 26) felt too much personal information is shared. Additional dislikes included the negative nature of some posts (n = 17; 15%), political posts (n = 11; 9%), and other responses that were unrelated (n = 52; 44%).

Finally, the question about privacy concerns generated the fewest responses. Of those responding, the biggest concern with privacy was the discomfort associated with the fact that people know things about users that users may not want shared (n = 81; 72%). The second most common concern was a fear of hacking or stealing of private information (n = 26; 23%). Finally, a small group answered the question by stating what they did to protect privacy or giving advice (n = 5; 5%).

Discussion

This is one of the first substudies of a randomized clinical trial to understand the role of social media for caregivers of hospice cancer patients. As expected, there was a significant difference in digital literacy between those who consented to the clinical trial and those who declined participation in the larger trial and only consented to the substudy for all aspects of digital literacy except social media usage. In short, those with greater digital literacy were generally more likely to consent to participate in the larger trial. Participants' responses to questions about social media usage (e.g., Do you check your [social media] account at least daily?) were not significantly different, and we suspect this was because usage was conditional on them using social media at all.

Although the participant-generated benefits of social media were easily grouped into three categories, it is interesting that the reasons family caregivers disliked social media varied so dramatically they could not be easily categorized. Clearly, many individuals appreciate the connectedness gained with social media, whereas some have concerns regarding others' provision of overly personal, political, or negative information, or wish to avoid the public sharing of too much of their own information. Privacy remains an ethical concern for much of the public and more data are needed to better understand its effect on the recruitment of patients and their family caregivers.

The negative association between age and digital literacy has implications for social media clinical trials. It is appropriate that when we recruit older persons, we are especially sensitive to their digital literacy levels and ensure that they have adequate assistance and information to make an informed consent decision and feel confident that, if they participate, they will be given the technical support required for the intervention so as not to cause additional burden or an unexpected privacy breach. Given the variance in digital literacy, these results suggest the importance of tailoring an informed consent process based upon digital literacy as a standard in recruitment to social media trials.

One of our research questions asked what is a brief way to assess individual digital literacy before an individual's decision to participate in a trial? We successfully identified two questions that allow us to do this. Asking potential social media trial participants to rate their skill with technology devices and their skill in the use of social media may be a helpful screen before consenting them into a social media study. Our analysis suggests that if someone responded with a rating of 5 or lower on technological skill or 4 or lower on social media skill (the means of nonenrolled for each), they should be given an enhanced recruitment protocol that recognizes their need for additional support and information before consenting.

These data lend insight in how we need to modify our intervention when we complete the trial and pursue the dissemination and implementation of the intervention. In addition, the data lend lessons for others doing technology-based interventions. In technology-based intervention studies, we would suggest that participants rating their skills with either the technological device or with the social media platform understand that with consent they would be offered additional services such as technology coaching, written information on how to use the social media, and perhaps video demonstration on how to use the media.

This customized recruitment and consent protocol could result in additional access to the clinical trial for those with limited digital literacy, a more informed participant as they are educated to the terminology and social media information, and a more representative sample of study participants.

To maximize the reach of promising digital health technologies in cancer care, it is imperative that we explore ways to engage patients and families with varying degrees of digital literacy in the research and testing of these new approaches. As noted in a recent article by Valdez et al.²⁷ it is important that with the emergence of new telehealth interventions, such as ACCESS, we become sensitive to the disparities and disabilities of those that cannot benefit from the opportunity and assure our interventions are inclusive of those with limited literacy and skill in their use.

LIMITATIONS

There are inherent biases within the results given the study was limited to caregivers of hospice cancer patients and had a relatively small sample. Although the clinical trial participants were randomized by hospice agency, the participants in this study were not randomized. Furthermore, to avoid overly burdensome onboarding processes given the complex nature of the study design, potential participants were only asked to provide consent to participate in the processes involved in the study arm (or group) to which their hospice agency was randomized at the time. It is, therefore, possible that digital literacy differentially affected family caregivers' willingness to participate in different study arms/groups. Given these limitations, this should be considered a pilot study with findings that indicate the need for further research.

Conclusions

Recruitment remains a challenge with seriously ill patient populations and their family caregivers, especially in hospice clinical trials. Finding new ways to support informed decision-making is critically needed as many factors may influence one's understanding and participation in research. Narrowing the gaps in digital literacy is one way to begin to explore the value of information technology and how it can help to tailor informed consent to specific contexts while enhancing participation in an ethical way that fully respects the rights of participants.

Authors' Contributions

All authors meet the criteria defined by the International Committee of Medical Journal Editors.

Disclaimer

The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Disclosure Statement

No competing financial interests exist.

Funding Information

Research reported in this publication was supported by the National Cancer Institute under award number 7R01CA203999 (P.O.).