Facebook Ads Recruit Parents of Children with Cancer for an Online Survey of Web-Based Research Preferences

Abstract

Background

Studies involving samples of children with life-threatening illnesses and their families face significant challenges, including inadequate sample sizes and limited diversity. Social media recruitment and web-based research methods may help address such challenges yet have not been explored in pediatric cancer populations.

Objective

This study examined the feasibility of using Facebook ads to recruit parent caregivers of children and teens with cancer. We also explored the feasibility of web-based video recording in pediatric palliative care populations by surveying parents of children with cancer regarding (a) their preferences for research methods and (b) technological capabilities of their computers and phones.

Methods

Facebook's paid advertising program was used to recruit parent caregivers of children currently living with cancer to complete an electronic survey about research preferences and technological capabilities.

Results

The advertising campaign generated 3,897,981 impressions which resulted in 1050 clicks at a total cost of $1129.88. Of 284 screened individuals, 106 were eligible. Forty-five caregivers of children with cancer completed the entire electronic survey. Parents preferred and had technological capabilities for web-based and electronic research methods. Participant survey responses are reported.

Conclusion

Facebook was a useful, cost-effective method to recruit a diverse sample of parent caregivers of children with cancer. Web-based video recording and data collection may be feasible and desirable in samples of children with cancer and their families.

Implications for Practice

Web-based methods (e.g., Facebook, Skype) may enhance communication and access between nurses and pediatric oncology patients and their families.

More than 11,500 children are diagnosed with cancer each year in the United States.¹ Despite medical advances, an estimated 1,350 children younger than 15 years old will die from cancer in 2014.¹ Research is crucial to improve the lives of children living with and dying from life-threatening conditions such as cancer, but recruitment challenges serve as potential study threats. Pediatric oncology and palliative care studies often have slower than expected accrual,^2,3 small sample sizes,^4,5 and limited generalizability. Further, studies involving face-to-face interaction between staff and participants during recruitment, intervention delivery, or data collection phases can contribute to participant burden and high study costs. Pediatric patients decline study because of not wanting to talk with health care professionals, lack of interest, or other engagements.⁶ Web-based methods for capturing and producing video-based interventions and data collection may decrease participant burden and appeal to children and teens. However, few studies have examined parent caregivers’ research preferences regarding web-based data collection or intervention delivery methods or the capabilities of their phones or computers to perform innovative, technologically advanced research activities such as video recording via distance. More work is needed in pediatric palliative care to better understand (a) how to add diversity to study samples to increase generalizability of results, (b) how to increase sample sizes to generate more rigorous analyses, (c) parent caregivers’ preferences for research methods, and (d) capabilities of parent caregivers’ phones and computers.

Social media recruitment may be one strategy to recruit larger and more diverse samples within pediatric palliative care research. Facebook ads have been used as a cost-effective and time-efficient recruitment method. Ramo and Prochaska⁷ obtained more than 3000 eligible consents by recruiting young adults who smoke across 13 months using Facebook ads. Facebook ads have successfully been used to recruit a variety of cohorts including gay and lesbian populations,^8,9 cigarette users,⁷ women of low income,¹⁰ college students,^11,12 and adults who are depressed.¹³ While the majority of studies have targeted adult age groups, Facebook ads have been used to recruit teenage and young adult populations such as Australian16 to 25 year olds¹⁴ and Canadian youth 15 to 24 year old affected by violence.¹⁵ One study using such ads involved recruitment of parents of children with Klinefelter syndrome.¹⁶ Facebook ads have yet to be used to recruit cancer patients of any age or their family members.

With Facebook's 1,310,000,000 active monthly users, 680,000,000 mobile users, and availability in 70 languages,¹⁷ Facebook ads may be a useful tool to identify parent caregivers of children eligible for pediatric palliative care and oncology studies. The feasibility and appropriateness of using Facebook recruitment for children with advanced cancer and their parent caregivers needs to be established. Web-based video recording may improve participant access, interest, and convenience; the feasibility of using such methods with pediatric populations has not been explored.

Study Objectives

The overall purpose of this study was to examine the feasibility of using social media recruitment methodologies via Facebook ads for parent caregivers (ages 18 years and up) of children and teens with cancer. Secondary aims of the study were to explore the feasibility of web-based video recording in pediatric palliative care populations by surveying parent caregivers of children with cancer regarding their preferences for survey and video recording research methods and capabilities of their computers and phones.

Methods

Facebook Advertising

Facebook's paid advertising program was used to recruit parent caregivers of children with cancer to complete an electronic survey about research preferences and technological capabilities. Participants were 18 years of age or up and legal guardians of children currently living with cancer. The ad ran from June 19, 2012 to August 31, 2012. The ad targeted age (18 years and up), location (United States), language (English), and childhood cancer related interest terms within Facebook profiles (e.g., childhood cancer, fight against childhood cancer, child cancer foundation, cure childhood cancer, pediatric cancer research foundation, pediatric oncology group, children's oncology group). Using Ramo and Prochaska's⁷ Facebook ad as a template, our ad contained a three-word title, photograph (which changed weekly), one-sentence description inviting parents of children with cancer to complete a short survey for a chance to win a $100 Apple gift card, and link to an electronic REDcap survey (Figure 1). REDCap is a secure, web-based application for building and managing online surveys and databases. Facebook approved the ad and specified ad size and word count. We determined a maximum lifetime total spending amount of $3000 for the ad and chose to pay per click using Facebook's suggested bid price of $0.34 to $0.66. Facebook determined the bid range based on how much other advertisers are willing to pay to reach the same target audience. Facebook automatically provided statistics on impressions (number of times an ad is shown to a user), clicks (when a user clicks on the ad link), and dollars spent.

Figure 1.

Example of Facebook Ad

Study Procedures

The ad was purchased after Institutional Review Board (IRB) approval. Participants who clicked on the survey link were directed to the REDCap survey cover page that described the study and explained that consent was shown by completion of the survey online. The first two questions screened potential participants for eligibility by asking if (1) they were 18 years of age or up, and (2) they were legal guardians of children 7 to 17 years of age who currently had cancer. For individuals answering “no” to either screening question, REDCap displayed a message that they were not eligible for study participation, and the survey concluded. Eligible participants who answered “yes” to both screening questions were directed to complete the survey about participant characteristics, child characteristics, research preferences, and phone and computer capabilities. The last question of the survey asked participants to provide their name, mailing address, and phone number to be entered into the drawing for a $100 Apple gift card. REDCap generated descriptive counts and frequencies for survey responses.

Results

Advertising Campaign

The advertising campaign generated 3,897,981 impressions (total number of times ad was shown on the site) which resulted in 1050 clicks (total number of clicks ad received), a 0.027% click through rate (number of clicks divided by the number of times ad was shown), and a total cost of $1129.88. The average cost per click was $1.08. Of 300 individuals who clicked on the survey, 106 were eligible. Of those 106 eligible, 67 participants started the survey and completed questions on participant and child characteristics. Of those 67 participants, 45 completed the entire survey (answering questions on participant and child characteristics, preferences for research, and phone/computer capabilities) (Figure 2).

Figure 2.

Facebook ad reach and recruitment

Participant and Child Characteristics

Of the 67 participants who completed survey questions regarding parent and child characteristics, the majority of participants were female (n = 52; 78%), located in the Southeastern region of the United States (n = 27, 40%), married (n = 48; 72%), biological parents (n = 62; 93%), and college undergraduates (n = 29; 43%). Forty percent (n = 27) of families had annual incomes between $25,000 to $50,000. Parents reported that their children with cancer averaged 10 years of age (SD = 4.40) and were mostly White (n = 59; 88%). Twenty-six (40%) of the children had experienced a cancer relapse or recurrence. Thirteen (20%) had palliative care or hospice teams involved in their care. See Tables 1 and 2 for participant and child characteristics.

Table 1.

Parent Caregiver Characteristics

Parent Caregiver Characteristics (N = 67)
		n (%)
Gender	Mother (female)	52 (77.6)
	Father (male)	15 (22.4)
Geographic region of the United States	Southeast (Alabama, Arkansas, Florida, Georgia, Kentucky, Louisiana, Mississippi, North Carolina, South Carolina, Tennessee, Virginia, West Virginia)	27 (40.3)
	Midwest (Illinois, Indiana, Iowa, Kansas, Michigan, Minnesota, Missouri, Nebraska, North Dakota, Ohio, South Dakota, Wisconsin)	11 (16.4)
	Southwest (Arizona, New Mexico, Oklahoma, Texas)	11 (16.4)
	Northeast (Connecticut, Delaware, Maine, Maryland, Massachusetts, New Hampshire, New Jersey, New York, Pennsylvania, Rhode Island, Vermont)	10 (14.9)
	West (Alaska, California, Colorado, Hawaii, Idaho, Montana, Nevada, Oregon, Utah, Washington, Wyoming)	8 (11.9)
Marital status	Married	48 (71.6)
	Divorced	10 (14.9)
	Single	5 (7.5)
	Separated	2 (3.0)
	Remarried	2 (3.0)
Highest grade level completed	Undergraduate	29 (43.3)
	High school	26 (38.8)
	Graduate school	9 (13.4)
	GED	3 (4.5)
Family income	$25,001 - $50,000	27 (40.3)
	$100,001 - or more	12 (17.9)
	$50,001 - $75,000	11 (16.4)
	Under - $25,000	10 (14.9)
	$75,001 - $100,000	7 (10.4)
Relationship to child	Biological parent	62 (92.5)
	Adoptive parent	4 (6.0)
	Other	1 (1.5)

Table 2.

Child Characteristics Per Parent Self-report

Child Characteristics (N = 67)
	Mean (SD)	Range
Age in years	10.09 (4.40)	1-17
Months since diagnosis	41.78 (37.17)	2-206
		n (%)
Race	White	59 (88.1)
	Other	7 (10.4)
	American Indian/Native American	1 (1.5)
Type of cancer	Solid tumors (e.g., sarcomas, neuroblastoma)	24 (39.3)
	Hematogenous malignancies (e.g., leukemia)	23 (37.7)
	Central nervous sytem tumors (e.g., medulloblastoma)	14 (23.0)
Cancer relapse or recurrence	Yes	26 (40.0)
	No	39 (60.0)
Palliative care or hospice team	No	52 (80.0)
	Yes	13 (20.0)

Research Preferences and Phone/Computer Capabilities

Forty-five participants answered survey questions regarding their research preferences and phone/computer capabilities. Ninety percent (n = 40) accessed the internet via personal computers. The majority of participants’ computers had broadband connections (n = 39, 87%), webcam or video capabilities (n = 33, 73%), speakers (n = 42, 93%), and microphones (n = 28, 62%). Sixty-seven percent (n = 30) knew how to use video chat (e.g., Skype/Facetime), and the majority (n = 10, 67%) who did not were willing to learn. Sixty-four percent (n = 29) already had video chat on their computer or phone, and the majority (n = 11, 69%) of those who did not were willing to download a free program. Eighty percent (n = 36) were willing to participate in a recorded web or phone interview. Sixty-nine percent (n = 31) were willing to assist their child in a recorded web or phone interview. Sixty-three percent (n = 26) preferred video recording for research via a scheduled web or phone video with a research nurse compared to videotaping on their own with a free loaned videocamera that they would return in a pre-paid envelope. Eighty-six percent (n = 30) of parents would allow audio and video recording of their child.

Discussion

This study leveraged social media research methodologies as a strategy to recruit parent caregivers of children with cancer. Overall, the Facebook ad was a successful, time efficient, and cost-effective approach to recruit a diverse sample of parent caregivers of children with cancer.

In 10.5 weeks, we recruited 67 parent caregivers of children with cancer. Forty percent (n = 26) of those parents reported that their child had experienced a relapse or recurrence. In contrast, our previous study took 12 months to recruit only 8 children (7 to 12 years of age) with cancer who had a poor prognosis (less than 60% chance of survival);² following a study modification, we recruited 28 children in 1 year (78% participation rate) by expanding to include children 7 to 17 years of age with any cancer at any phase of treatment, including new diagnoses or relapses. Results suggest that Facebook ad recruitment strategies may result in expedited recruitment of eligible participants compared to traditional recruitment methods. Facebook recruitment may benefit multiple types of research involving human subjects (e.g., randomized controlled trials, qualitative) and be particularly useful for studies targeting rare populations, such as children with cancer who have a poor prognosis, relapsed or recurrent cancer, or refractory disease.

Our findings also suggest that Facebook recruitment may efficiently enable researchers to increase sample diversity and subsequent generalizability of study results. Our participants represented all regions of the United States and varied across family income and education levels. Twenty-two percent of our participants were Fathers, a historically underrepresented participant population.¹⁸ Parent caregivers reported that their children with cancer averaged 10 years of age, which was similar to the average age (M = 11) of children with cancer recruited in person in a recent study at our own institution.² Similarly, race, education, and relationship to child were comparable to previous samples recruited via traditional methods in our previous single-site² and multi-site studies.¹⁹ Ultimately, Facebook recruitment captured a sample with similar demographic characteristics to previous samples but may be a more cost-effective approach, generating greater geographic diversity and improved generalizability of results compared to single-site or multi-site studies using traditional recruitment methods.

Interestingly, we paid $1129.88 to advertise on Facebook which is less than $17 per participant. Facebook recruitment required minimal staff and may be much more cost-effective compared to paying a trained research assistant or project director to approach potential participants via traditional in-person or telephone methods. Face-to-face recruitment can involve extensive waiting time before or after clinic visits to coordinate with family members’ schedules.² Similarly, telephone recruitment can often take several hours per participant since numerous phone calls are often required to contact and recruit family members of children with cancer.¹⁹

Results indicated that web-based video recording and data collection are likely feasible and preferred by some parents of children living with cancer compared to guided self-video recording methods (e.g., with a Flip video camera) and mailed paper surveys. The majority of caregivers had phones or computers capable of supporting web-based surveys and video recordings. Parent caregivers that did not have these means reported that they were generally willing to download appropriate programs and accept assistance to support their participation. The majority of parents reported that they would support their child's participation and assist them with video recordings and surveys. These findings support high rates of internet access, similar to the 2011 United States Census Bureau that reported more than 80% of Americans ages 3 years and older live in a home with at least 1 computer and 27% access the internet from both inside and outside the home with multiple devices (e.g., computer, Smartphone).²⁰ Researchers should still budget for equipment (e.g., Smartphones with minutes of usage) in future grants for participants who may not have a phone or computer access. Further research should examine the feasibility of web-based video recording with pediatric populations. Additionally, comparing outcomes, overall expense, and time efficiency of studies using such methods (e.g., social media recruitment, web-based data collection and video recording) to studies using more traditional research methods (e.g., telephone or in-person recruitment, face-to-face video recording) is needed.

Limitations

We did not track parents from the same family who may have completed the REDCap survey; however, this has been done in other studies⁷ and could be incorporated into future work. Although our clinical experience supports that parent caregivers can accurately self-report information regarding their child's illness, we were unable to evaluate accuracy of parents’ self-reports of children's illness characteristics (e.g., type of cancer, months since diagnosis) since we did not have access to medical records. A future evaluation comparing parent self-reports to medical chart reviews could provide useful information to institutional review boards (IRB) that may question this access and accuracy issue. We only recruited across 10.5 weeks which limited our sample size, yet this data collection time period was appropriate for our study aims. Finally, only 33% of participants partially completed the REDCap survey. We suspect that this resulted from participants accidentally failing to click “submit” at the bottom of the survey page which was required to advance to the next set of survey questions. Future studies would likely decrease missing data by formatting surveys to automatically continue to the next question, if possible. Researchers may also decrease missing data by developing a protocol for missing data. For example, a staff member could follow-up via email with participants who submitted partially completed responses, asking them to respond to questions they may have accidentally skipped.

Obtaining IRB Approval

Although we did not have challenges with getting IRB approval from our institution, new recruitment methods, particularly with vulnerable populations such as children with life-threatening conditions, may challenge some IRBs. Strategies that may have helped us obtain approval included talking with an IRB representative before submitting the application to clarify issues that need to be addressed, describing the recruitment method in clear detail in the IRB application, use of our secure web-based data collection program (REDCap) to help maintain data confidentiality, and inclusion of a REDCap introduction to participants that clearly described study details and described the voluntary nature of the study. Researchers encountering IRB challenges could provide references to IRB reviewers, such as this article, to provide support for the method being used in other studies without harm to participants.

Conclusion

Web-based recruitment and data collection methods were useful and cost-effective ways to recruit and obtain self-reports from a diverse sample of parent caregivers of children with cancer. Web-based methods such as Facebook or Skype may be effective tools for nurses and other clinicians to communicate with parents of children with cancer; this could be especially useful for families who may live in rural areas or long distances from the hospital. Future research could compare psychosocial variables (e.g., quality of life, coping) between samples recruited via traditional and social media recruitment methods or examine feasibility of collecting longitudinal web-based data via samples recruited via social media recruitment methods. More research is needed to develop new web-based recruitment and data collection methods that are feasible, cost-effective, acceptable to pediatric oncology and palliative care populations, and have the potential to recruit larger, more diverse samples. Knowing one's research population is vital to determining the most effective recruitment and data collection strategies. Scientists can involve patients early in the research process (e.g., conduct preliminary focus groups, pilot recruitment method studies) to refine new research methods and individualize recruitment and data collection approaches for target populations. Researchers, nurses and other clinicians, and review boards should be contemporary thinkers and receptive to consider innovative ideas as new evidence-based methods could springboard the cancer and palliative care field. Results will provide crucial information to improve sample diversity, increase sample sizes, improve generalizability of results, and ultimately enhance rigor in pediatric palliative care and oncology research.

Table 3.

Participant (N = 45) Research Preferences and Phone/Computer Capabilities

Parent Caregiver Research Preferences
		n (%)
I prefer to complete surveys for research via the following methods:	Private and secure online web-based survey	40 (88.9)
	Paper copy survey received in the mail that I would return in a pre-addressed stamped envelope provided to me	5 (11.1)
I would be willing to participate in a recorded web or phone interview with a research nurse.	Yes	36 (80.0)
	No	9 (20.0)
I would be willing to assist my child in an interview via web/phone video materials such as Skype or Facetime.	Yes	31 (68.9)
	No	14 (31.1)
I prefer videotaping for research via the following methods:	Scheduled web/phone video with a research nurse (e.g., via Skype or Facetime)	26 (63.4)
	With guidance regarding what to record or talk about, I prefer to videotape on my own with a videocamera (e.g., Flip video camera) loaned to me for free that I would mail back in a prepaid envelope	15 (36.6)
I would be willing to assist my child in videotaping with a provided Flip video camera	Yes	32 (71.1)
	No	13 (28.9)
I would agree to the following types of recording for my child:	Audio and video recording (recording your child's voice and face)	30 (85.7)
	Audio recording (recording your child's voice)	14 (40.0)
Phone/Computer Capabilities
How do you access the internet?	Personal computer/PC (such as Dell)	40 (93.0)
	SmartPhone	18 (41.9)
	Mac computer	9 (20.9)
My computer has a broadband connection (e.g., cable, DSL).	Yes	39 (86.7)
	No	5 (11.1)
	Unsure	1 (2.2)
My computer has a webcam/video capability.	Yes	33 (73.3)
	No	9 (20.0)
	Unsure	3 (6.7)
My computer has speakers so that I can hear sound.	Yes	42 (93.3)
	No	3 (6.7)
My computer has a microphone.	Yes	28 (62.2)
	No	13 (28.9)
	Unsure	4 (8.9)
I have Smartphone acess with face-to-face video chat capability.	Yes	20 (44.4)
	No	24 (53.3)
	Unsure	1 (2.2)
I currently know how to use video chat such as Skype or Facetime.	Yes	30 (66.7)
	No	15 (33.3)
I currently have video chat access on my computer or phone, such as Skype or Facetime.	Yes	29 (64.4)
	No	12 (26.7)
	Unsure	4 (8.9)
How often do you use video chat such as Skype or Facetime	Only once or a few times	12 (41.4)
	Frequent user	7 (24.1)
	Once a month	6 (20.7)
	Several times a year	4 (13.8)
What video chat program do you have?	Skype	20 (69.0)
	Facetime	7 (24.1)
	Other	2 (6.9)