Abstract
Recruiting pregnant smokers into clinical trials is challenging since this population tends to be disadvantaged, the behavior is stigmatized, and the intervention window is limited. The purpose of this study is to test the feasibility and effectiveness of recruiting pregnant smokers into a smoking cessation trial by sending recruitment text messages to an existing subscriber list. Recruitment messages were sent to subscribers flagged as pregnant in Text4baby, a national text messaging program for pregnant women and mothers. Four recruitment messages were rotated to test the effectiveness of different emotional frames and a financial incentive. Study staff called subscribers who expressed interest to screen for eligibility and enroll eligible women. Between October 6, 2015 and February 2, 2016, 10,194 recruitment messages were sent to Text4baby subscribers flagged as pregnant, and 10.18% (1038) responded indicating interest. No significant increase in cancellation was observed compared to subscribers who received other ad hoc messages. Of respondents, 54.05% (561) were reached by phone for follow-up, and 21.97% (228) were found to be eligible. Among the eligible, 87% (199) pregnant smokers enrolled. The recruitment message with a pride emotional appeal had a significantly higher response (p = 0.02) compared to the recruitment message with no emotional appeal, but enrollment did not significantly differ between recruitment messages with different emotional appeals. The recruitment messages with a reference to financial incentive yielded higher response (p < 0.01) and enrollment (p = 0.03) compared to a recruitment message without. This study demonstrates success recruiting pregnant smokers using text message. Future studies should consider building on this approach for recruiting high-risk populations.
Keywords: mHealth, Pregnant women, Smoking cessation, Quit4baby, Text4baby
BACKGROUND
Smoking during pregnancy has been shown to cause adverse fetal outcomes such as low birth weight, preterm birth, fetal growth restriction, and pregnancy complications such as ectopic pregnancies [1], in addition to the risks generally associated with smoking. It is estimated that at least 5% of preterm-related deaths, at least 23% of sudden infant deaths [2], and 13.1–19.0% of term low birth weight deliveries could be prevented by eliminating smoking during pregnancy [1].
Despite these known health risks, only 24.2% of women who smoked in the 3 months prior to pregnancy were able to quit before pregnancy [3] and an estimated 8.4–10.2% of women smoke during pregnancy in the USA [3, 4]. Furthermore, smoking rates during pregnancy are higher among women who are 20–24 years old (16.8%), among those with less than a high school education (14.1%) and among those with Medicaid insurance (14.0%) [3].
Pregnant women are underrepresented in clinical trials [5], and many of the clinical trials recruiting pregnant women have had difficulty meeting their enrollment goals [6–13]. Barriers to recruiting pregnant smokers into clinical trials include that this population tends to be socially disadvantaged [14], some may avoid or delay seeking prenatal care in the healthcare system [15], and some may be reluctant to disclose their smoking status to providers because the behavior is stigmatized [16, 17]. Another challenge of recruiting pregnant smokers is that the intervention window defined by pregnancy is limited.
Effective techniques for recruiting hard to reach populations like pregnant smokers include providing an incentive [18–21] and targeting recruitment messages and materials to the population [5, 22–24]. Using emotion to frame messages may be an effective form of targeting. Distinct emotional appeals (e.g., pride, empathy) may have differential impacts on health behaviors [25], and high affect-arousing messages may be particularly persuasive among pregnant smokers [26]. The effects of emotional appeals have long been studied [27], but most have solely focused on fear [28] which is, perhaps, an inappropriate emotion to evoke in pregnant women. Thus, the focus here is to examine unique emotional frames to assess which might be effective with this population.
Text messaging on mobile phones has become a ubiquitous communication channel among women of child-bearing age and may be an effective way to reach pregnant smokers. Among adults ages 18 to 29 years, 98% own a cell phone [29], and on average, they send and receive 87.7 texts per day [30]. Adults who are low income—which is typical of pregnant smokers—send and receive more texts compared to higher income adults [30]. Additionally, compared to other modalities (e.g., in-person, phone), text messaging may offer higher levels of anonymity and confidentiality [17].
Text4baby is the largest text messaging service for pregnant women and mothers in the USA with more than 1 million pregnant women and new mothers enrolled since its inception and more than 150,000 women enrolled a year [31]. The service sends three texts per week to women flagged as pregnant that are timed to her due date, as well as modules of messages on health issues such as influenza vaccination and ad hoc alert messages with breaking news on disease outbreaks [32]. Text4baby is promoted by a large network of more than 1400 partners including government agencies, such as state and local health departments and Medicaid agencies, health plans, professional associations, and non-profit organizations [31, 33, 34]. According to a survey of Text4baby subscribers, subscribers represent underserved populations with the majority reporting being covered by Medicaid and having an annual household income of less than $16,000 [35].
PURPOSE
The purpose of this study was to build on an earlier pilot [36] and to determine the feasibility and effectiveness of recruiting pregnant smokers from the Text4baby subscriber base for an efficacy trial of Quit4baby. Quit4baby is a smoking cessation text messaging program for pregnant women envisioned as an add-on service for smokers in Text4baby, as well as a standalone smoking cessation program [37]. While the Text4baby subscriber list has previously been used to recruit for Text4baby program evaluation efforts [29, 34], this study was the first to recruit via text message for a separate trial and to study the recruitment process. This paper does not report results from the efficacy trial rather it aims to contribute to the practice-based evidence [38] for recruiting high-risk population by focusing on the feasibility and effectiveness of recruitment. A supplementary goal of the study was to understand if varying the content of the recruitment text message—by changing the message appeal and removing mention of a financial incentive—affected the response and enrollment into the trial.
METHODS
A recruitment text message was sent to Text4baby subscribers flagged as pregnant to identify interested pregnant smokers for the Quit4baby trial. The Quit4baby trial had an enrollment target of 500 participants. Between August 2015 and February 2016, Text4baby subscribers were sent a recruitment message if they were flagged as pregnant, had reported a due date at least 8 weeks in the future, and were actively enrolled in the program at the time the recruitment message was sent (i.e., had not unsubscribed from Text4baby). Subscribers who registered for Text4baby with a zip code from California, Oklahoma, Ohio, or Louisiana were excluded because the Quit4baby program was already available to Text4baby subscribers in those states. Recruitment messages were broadcasted at 2:00 p.m. Eastern Standard Time, so subscribers received the recruitment message between 11:00 a.m. and 2:00 p.m. local time (or 8:00 a.m. for Hawaii subscribers). Tuesday was selected because this assured that in most cases subscribers would only receive one message on this day given that Text4baby standard weekly messages are not sent on Tuesdays. Subscribers were grouped into new subscribers, those who had signed up for the Text4baby service 4 to 10 days prior to the sending of the recruitment message, and older subscribers, those who had been in the program for longer than 10 days. For new subscribers, the 4 to 10 day window was chosen to give Text4baby subscribers a few days to become acclimated with the program before presenting them with the study offer and allowed the study team to send the broadcast on a weekly basis without sending the recruitment message to the same Text4baby subscriber twice.
Text4baby subscribers who responded “YES” to the recruitment message were then followed-up with by phone to complete the enrollment process. Research staff called interested Text4baby subscribers at least four times within the first week and up to eight times if a respondent continued to express interest until the Text4baby subscriber verbally refused or was screened for eligibility. Research staff assessed participants for eligibility by conducting a 3 to 5 min telephone screener. Participants were eligible for the Quit4baby study if they had a cell phone for their personal use, were willing to receive text messages on their cell phone, were 14 years or older, were currently pregnant, and had smoked at least one puff of a cigarette in the past 2 weeks. The number of follow-up attempts it took to enroll participants was tracked. If eligible and interested, Text4baby subscribers consented to participate in the Quit4baby trial and were enrolled over the phone. The recruitment protocol was approved by the George Washington University Institutional Review Board.
For purposes of this analysis, the recruitment efforts that targeted only new subscribers between October 2015 and February 2016 are reported here. New subscribers were expected to be more responsive to a recruitment text than older subscribers who had been in the program for a longer period of time and had become more accustomed to receiving Text4baby appeals. Thus, new subscribers were selected because they were more uniform and the effects of the recruitment efforts could be better isolated. The data from the months of August and September 2015 were not included because the recruitment procedures and message content were being refined during this time.
Recruitment text messages
Recruitment text messages aimed to identify pregnant smokers within Text4baby who would be willing to be part of a smoking cessation study. All texts had the same call to action, asking subscribers to reply “YES” to learn more. Three of the messages varied the central appeal of the message, with two aimed at an emotional appeal (i.e., pride, empathy) and one emphasizing a free service. The pride message emphasized feelings of pride from quitting. This message was as follows (with italics for emphasis of the appeal): “Text4baby: If you smoke, it’s important to quit. You’ll be proud you did! Test messages to help you quit. Get gift cards if eligible. Reply YES to learn more.” The empathetic message recognized the difficulty in quitting smoking and offered social support. This message was as follows: “Text4baby: Quitting smoking is tough but it’s easier with help. Test messages to help you quit. Get gift cards if eligible. Reply YES to learn more.” The third message conveyed that the service was free: “Text4baby: Want FREE messages to help you quit smoking? Test messages to help you quit. Get gift cards if eligible. Reply YES to learn more.” One additional variant of the pride message excluded the reference to the gift card incentive in order to test whether the mention of a financial incentive drove participation. The gift card reference—“Get gift cards if eligible” was removed, as follows: “Text4baby: If you smoke, it’s important to quit. You’ll be proud you did! Test messages to help you quit. Reply YES to learn more.”
Research staff planned to rotate the three types of appeals every Tuesday; however, the rotation was not practical to maintain given multiple factors, such as holidays. From October 2015 to February 2, 2016, the recruitment message was sent to new subscribers on 12 Tuesdays; the pride message was broadcasted 3 times, the free message was broadcasted 3 times, and the empathy message was broadcasted 5 times. Additionally, the pride message without the written reference to the incentive present was broadcasted once on February 9, 2016. This message was compared to the pride message with the written reference to the incentive only and was not included in the overall analysis unless specifically stated. The total number of recruitment messages varied based on how many weeks each message type was sent and the amount of subscribers who met the selection criteria each week.
Measures and analysis
Feasibility of recruitment was defined as the ability to send recruitment text messages to Text4baby subscribers flagged as pregnant and to send the recruitment messages without causing a significant spike in cancellations compared to other ad hoc messages sent by Text4baby. Text4baby subscribers were able to cancel the Text4baby program at any time by texting the keyword STOP. The percent of subscribers who cancelled during the recruitment period was defined as the number of Text4baby subscribers who cancelled from the Text4baby program on the Tuesday the recruitment message was sent. Three of the 11 days (2791 subscribers) were excluded from the cancellation calculation given that on these days Text4baby subscribers also received a Text4baby ad hoc message or topical module message and cancellation associated with just the recruitment message could not be determined. The percent who cancelled on a recruitment Tuesday were compared to the percent who cancelled among a similar group of pregnant, newly enrolled, Text4baby subscribers who received a Text4baby ad hoc message or topical module message on a Tuesday during the previous 6 months (between January 6, 2015 and July 14, 2015). These measures will test the operational feasibility of using the Text4baby platform for recruitment without disrupting the existing service.
The effectiveness of the recruitment efforts was measured in three main ways. First, effectiveness was measured by the percentage of subscribers who expressed interest via text. Percent interested was calculated by the number of subscribers who replied YES to the recruitment message out of the total number of subscribers who were sent the recruitment message. Second, effectiveness was measured by the percentage of subscribers who enrolled. Percent enrolled was calculated by the number enrolled out of the total number of subscribers sent the recruitment message. Third, in order to understand what fraction of the targeted population was reached, an important evaluation measure [39], effectiveness was measured by the percent of estimated pregnant smokers who enrolled. This was calculated by the number enrolled out of the total number of pregnant Text4baby subscribers who were estimated to be smoking based on the national prevalence rate of 10.2% [4]. This rate was considered a conservative estimate for pregnant Text4baby subscribers who represent low-income populations [31, 37] and are more likely to be smoking [3]. Other measures of effectiveness include percent contacted by phone, percent verbally refused, and the number of ineligible and eligible subscribers who were sent the recruitment message.
Text4baby subscribers’ response to the recruitment message and timing of text response was captured by Text4baby system data. The number of follow-up attempts it took to enroll interested and eligible subscribers by phone was tracked by research staff within the recruitment database. The eligibility screener, consent process, and baseline questionnaire were managed using REDCap electronic data capture tools hosted at the Children’s National Medical Center [40]. The baseline questionnaire captured participant socioeconomic characteristics, which were subsequently used to determine representativeness across subgroups.
The measures of recruitment feasibility and effectiveness were compared across the different messages using Fisher’s exact test for significance with a 95% confidence level. A logistic regression was run to measure the odds of enrollment based on time to response to the recruitment messages. SAS 9.4 was used to conduct all statistical analyses.
RESULTS
Between October 6, 2015 and February 2, 2016, recruitment messages were sent to 10,914 Text4baby subscribers flagged as pregnant (see Table 1 for overview of feasibility and effectiveness results). The percent of Text4baby subscribers who were sent the recruitment message and cancelled was 1.78%. Percent cancelled during the comparison period for subscribers who received a Text4baby ad hoc or topical module message was 1.86% (64 of 3439). There was no statistical difference detected in cancellation between both groups.
Table 1.
Overview of feasibility and effectiveness of recruiting through text
| Feasibility | Effectiveness | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Date | Message type | messages sent | Cancelleda | % Cancelled (95% CI)b | Interested | % Interested (95% CI)c | Enrolled | % Enrolled (95% CI)d | % of smokers enrolled (estimate)e |
| October 6, 2015 | Free | 886 | 22 | 2.48% (1.56–3.74%) | 74 | 8.35% (6.61–10.37%) | 14 | 1.58% (0.87–2.64%) | 15.49% |
| October 20, 2015 | Empathy | 974 | 12 | 1.23% (0.64–2.14%) | 121 | 12.42% (10.42–14.66%) | 24 | 2.46% (1.59–3.64%) | 24.16% |
| October 27, 2015 | Empathy | 847 | 9 | 1.06% (0.49–2.01% | 93 | 10.98% (8.95–13.28%) | 17 | 2.01% (1.17–3.19%) | 19.68% |
| November 3, 2015 | Pride | 888 | – | – | 109 | 12.27% (10.19–14.62%) | 21 | 2.36% (1.47–3.59%) | 23.18% |
| November 10, 2015 | Free | 836 | 25 | 2.99% (1.94–4.38%) | 80 | 9.57% (7.66–11.77%) | 16 | 1.91% (1.10–3.09%) | 18.76% |
| November 17, 2015 | Empathy | 1147 | 18 | 1.57% (0.93–2.47%) | 109 | 9.5% (7.87–11.35%) | 15 | 1.31% (0.73–2.15%) | 12.82% |
| December 8, 2015 | Empathy | 1118 | – | – | 101 | 9.03% (7.42–10.87%) | 25 | 2.24% (1.45–3.28%) | 21.92% |
| January 12, 2016 | Pride | 871 | 11 | 1.26% (0.63–2.25%) | 91 | 10.45% (8.50–12.67%) | 18 | 2.07% (1.23–3.25%) | 20.26% |
| January 19, 2016 | Free | 858 | 13 | 1.52% (0.81–2.58%) | 77 | 8.97% (7.15–11.09%) | 16 | 1.86% (1.07–3.01%) | 18.28% |
| January 26, 2016 | Empathy | 785 | – | – | 81 | 10.32% (8.28–12.66%) | 16 | 2.04% (1.17–3.29%) | 19.98% |
| February 2, 2016 | Pride | 984 | 22 | 2.24% (1.41–3.37%) | 102 | 10.37% (8.53–12.44%) | 17 | 1.73% (1.01–2.75%) | 16.94% |
| Total | 10,194 | 132 | 1.78% (1.08–1.53%) | 1038 | 10.18% (9.60–10.79%) | 199 | 1.95% (1.69–2.24%) | 19.14% | |
aDays when Text4baby subscribers were sent another ad hoc or topical module message were excluded
bThe number of “STOP” text responses among the number of recruitment messages sent
cThe number of “YES” text responses among the number of recruitment messages sent
dThe number enrolled in the trial among the number of recruitment messages sent
eThe number enrolled in the trial among an estimated 10.2% of smokers who were sent a recruitment message
Of the 10,194 Text4baby subscribers who were sent a recruitment message, 1038 (10.18%) texted YES indicating interest. Of those (1038), almost half of the respondents (477, 45.95%) were not reachable after their initial text for several reasons including no answer to study calls, inability to receive incoming calls, and phone being out of service. Research staff reached 561 (54.05%) respondents via phone call for follow-up and 152 (14.64%) refused, 181 (17.44%) were screened and found to be ineligible, and 228 (21.97%) were screened and found to be eligible. Reasons for ineligibility included not smoking a cigarette within the past 2 weeks (168, 92.82%), not currently pregnant (7, 3.87%), and not having a cell phone for their personal use (1, 0.01%) or a texting plan on their cell phone (5, 2.76%). Among the ineligible due to smoking status, 84 (50.00%) smoked every once in a while or quit smoking once they found out they were pregnant, 24 (14.88%) used to be regular cigarette smokers, and 60 (38.10%) never were regular cigarette smokers. Of those who were assessed for eligibility and found to be eligible (228), 199 (87.28%) went on to enroll in the study (see Fig 1 for the recruitment diagram). The percent who enrolled among Text4baby subscribers who were sent the recruitment message was 1.95%. Based on the 10.2% smoking prevalence among pregnant women in the USA [4], it is estimated that 1040 subscribers who were sent the recruitment message were smoking and therefore eligible to participate in the Quit4baby trial. Thus, an estimated 19.14% (199/1040) of smokers in Text4baby who were sent the message were enrolled (see Table 1 for feasibility and effectiveness measures by week).
Fig 1.
Recruitment flowchart
The mean text response time to the recruitment message was 23 min (median is 212.05 min or 3.53 h) with response times ranging from 0 min to 12,873 min (8.84 days). There was no relationship between response time to the recruitment message and enrollment. The mean number of attempts to enroll was 2.35 (SD = 1.67) among participants. There were no significant differences among participants who enrolled on the first or second attempt compared to all other participants based on their education, race, or cigarettes smoked per day at baseline.
Effectiveness of recruitment message framing
Of the 10,194 subscribers who received a recruitment message, 2580 received the free message, 4871 received the empathetic message, and 2743 received the pride messages with reference to the incentive. The free message had a significantly higher cancellation (2.33%) compared to the empathy message (1.31%) (p < 0.01), but the pride message cancellation (1.78%) did not significantly differ from the free or empathy message cancellation. Subscribers responded significantly more to the pride message (10.97%) compared to the free message (8.99%) (p = 0.02). Subscribers responded more to the empathetic message (10.37%) compared to the free message (8.99%) (p = 0.06), but the results were not significant. There were significant differences in the proportion of subscribers who were ineligible between the three messages; the pride message had the most ineligible (2.44%), which was significantly more than the empathetic message (1.66%) (p = 0.02) and the free message (1.40%) (p < 0.01). Reasons for ineligibility did not significantly differ between the three groups. The proportion enrolled between recruitment messages was not significantly different. Thus, although the emotionally salient messages had a significantly higher response compared to the free message, the higher response did not translate to higher enrollment. The percent contacted via phone for follow-up and the percent refused after the initial text response also did not significantly differ between the three recruitment messages (See Table 2 for feasibility and effectiveness by message frame).
Table 2.
Overview of feasibility and effectiveness of recruitment by message frame
| Feasibility | Effectiveness | |||||||
|---|---|---|---|---|---|---|---|---|
| Message type | Messages sent | Cancelleda | % Cancelled (95% CI)b | Interested | % Interested (95% CI)c | Enrolled | % Enrolled (95% CI)d | % of smokers enrolled (estimate)e |
| Free | 2580 | 60 | 2.33% (1.78–2.98%) | 232 | 8.99% (7.92–10.16%) | 46 | 1.78% (1.31–2.37%) | 17.48% |
| Empathy | 4871 | 39 | 1.31% (0.94–1.79%) | 505 | 10.37% (9.53–11.26%) | 97 | 1.99% (1.62–2.42%) | 19.52% |
| Pride | 2743 | 33 | 1.78% (1.23–2.49%) | 301 | 10.97%f (9.83–12.20%) | 56 | 2.04% (1.55–2.64%) | 20.02% |
| Total | 10,194 | 132 | 1.78% (1.49–2.11%) | 1038 | 10.18% (9.60–10.79%) | 199 | 1.95% (1.69–2.24%) | 19.14% |
aDays when Text4baby subscribers were sent another ad hoc or topical module message were excluded. The total excludes November 3, 2015; December 8, 2015; and January 26, 2016
bThe number of “STOP” text responses among the number of recruitment messages sent on the days Text4baby subscribers received the recruitment message only (free messages sent = 2580, empathy messages sent = 2968, and pride messages sent = 1855; total n = 7403)
cThe number of “YES” text responses among the number of recruitment messages sent
dThe number enrolled in the trial among the number of recruitment messages sent
eThe number enrolled in the trial among an estimated 10.2% of smokers who were sent a recruitment message
fFisher’s exact test two-tailed p value is <0.05 where empathy and pride messages were compared to the free message
The pride message without reference to the gift card incentive was sent to 724 new Text4baby subscribers. The percent cancelled (1.24%) was not significantly different than the percent cancelled among subscribers who received the pride message with reference to the incentive (1.78%). The percent of subscribers who responded (7.18%) (p < 0.01), the percent of subscribers who enrolled (0.83%) (p = 0.03), and the percent of subscribers who were found to be ineligible (0.83%) (p < 0.01) were significantly lower compared to the pride message with reference to the incentive.
Table 3 provides an overview of the demographic characteristics of participants recruited through this method. Participants represent low income; more than half had an annual household income less than $15,000 and 81% had an annual household income less than $30,000. More than half of participants did not work at all (121, 60.80%) and had Medicaid insurance (148, 74.37%). On average, participants reported smoking 15.95 (SD = 10.17) cigarettes per day before finding out they were pregnant and 6.75 (SD = 5.87) cigarettes per day at the time of study enrollment. Participant’s race, education, employment status, insurance type, and cigarettes smoked per day at baseline did not significantly differ by recruitment messages.
Table 3.
Baseline demographic characteristics (N = 199)
| Frequency mean | %, SD | |
|---|---|---|
| Age | 26.33 | 6.15 |
| Racea | ||
| White | 133 | 66.83 |
| Black | 54 | 27.14 |
| Other | 15 | 7.54 |
| Hispanic, Latina, or Spanish origin | 17 | 8.54 |
| Incomeb | ||
| Up to $15,000 | 108 | 54.27 |
| $15,001–$30,000 | 53 | 26.63 |
| $30,001–$47,099 | 22 | 11.06 |
| $47,100 or more | 9 | 4.52 |
| Education | ||
| 12th grade or less with no high school diploma | 50 | 25.13 |
| High school graduate, GED or equivalent | 64 | 32.16 |
| High school graduate | 49 | 24.62 |
| Some college | 63 | 31.66 |
| Associate’s, bachelor’s, or master’s | 20 | 10.05 |
| Workc | ||
| Part time | 40 | 20.10 |
| Full time | 35 | 17.59 |
| Not at all | 121 | 60.80 |
| Insurance | ||
| None | 14 | 7.04 |
| Medicaid/medicare | 148 | 74.37 |
| Veterans or military | 21 | 10.55 |
| Private insurance/from employer | 14 | 7.04 |
| Had a smartphoned | 172 | 85.15 |
| Cigarettes smoked per day before pregnancy | 15.95 | 10.17 |
| Cigarettes smoked per day now | 6.75 | 5.87 |
| Weeks Pregnant | 16.24 weeks | 54.42 |
All variables had missing data for two participants unless otherwise stated
aNot mutually exclusive
bMissing data for seven participants
cMissing data for three participants
dMissing data for three participants
DISCUSSION
Recruitment for a smoking cessation trial through Text4baby was found to be operationally feasible and effective [39]. Within a 5-month period, more than 10,000 new Text4baby subscribers flagged as pregnant were sent a recruitment text message. The percent who cancelled as a result of the recruitment message was found to be similar to cancellation from other ad hoc messages sent to Text4baby subscribers. This suggests that this recruitment approach did not disrupt the Text4baby service more so than other ad hoc messages sent by the program.
More than 10% of Text4baby subscribers flagged as pregnant who were sent a recruitment message replied expressing interest in the study and given that about 10.2% of pregnant women smoke in the USA based on national averages [4], this recruitment approach was successful at obtaining interest from the target population of pregnant smokers receiving Text4baby. The success may be because of the conveniences associated with text messaging; the target audience could read the short solicitation on their own time without much additional effort and easily express interest (i.e., reply YES). Recruiting by text message from an existing text messaging subscriber list has a number of advantages. It was possible to rapidly identify potential study participants from a large, national pool of pregnant women on a weekly basis, perhaps using fewer study resources than if more traditional approaches were used such as online advertisements, provider outreach, or making phone calls to marketing lists of pregnant smokers [6]. Also, because text responses were received within minutes, the study team could follow-up with interested Text4baby subscribers close to their time of decision.
Although interest in the recruitment message was high, research staff had to persistently follow-up with interested subscribers’ multiple times to enroll them in the study, and almost half of subscribers who expressed interest via text never answered the follow-up phone call (477, 45.95%). Of the participants who took the phone call, many verbally refused or were ineligible (333, 59.36%). On average, it required more than two phone attempts to enroll an interested Text4baby subscriber in the trial. It is important to note that participants who enrolled on the first phone attempt were representative of the entire study population. Given this, for future studies of this nature, it may be possible to use staff time more efficiently by focusing on refining procedures for those who answer on the first phone attempt.
The recruitment message appeal was important for generating initial responses via text. Recruitment messages that had a pride emotional appeal had a better response over one that offered a free service. Furthermore, the proportion of subscribers found to be ineligible differed between the three groups. This may have been the case because non-smokers or women who recently quit sympathized with pregnant smokers more after receiving the pride or empathetic message and wanted to help. There is potential to increase enrollment in the future, perhaps by using other types of appeals (e.g., fear-based).
As expected, removing the reference to the incentive in the recruitment message significantly reduced response and enrollment. Nonetheless, it is noteworthy that a portion of subscribers (7.18%) still responded to the recruitment message without reference to the incentive suggesting that, while inferior, this recruitment message may still be effective. It may be that the emotionally salient message alone elicited a perception on non-tangible benefits, and these perceived benefits outweighed the cost of participation [6].
A limitation of this study was that the different messages were not randomly assigned at the individual level or randomized across weeks. Therefore, there is potential for confounding as a result of seasonality or other changes that may have occurred over time. Furthermore, the representativeness of Text4baby pregnant smokers compared to other pregnant smokers must be considered in generalizing the results of the study. Text4baby subscribers may differ in their comfort with texting and may be more willing to be recruited via text message compared with other pregnant smokers such as those who may be recruited through a health system.
Despite these limitations, the results herein support the utility of text messages for study recruitment, as well as the importance of testing message elements, including the type of appeal and the inclusion of incentives. The results may have implications not only for recruiting pregnant smokers into trials but also for existing smoking cessation services and demonstrates the importance of more practice-based research [38].
Acknowledgments
The authors would like to thank the George Washington University Research Assistants Dasha Afanaseva MPH, Laura Macherelli MPH, Shelby Fallon MPH, Shawn Chiang, and Nisha Radhakrishnan for following through with the recruitment process over the phone. We would also like to thank Ada Obi, Lalida Thaweethai MPH, Diana Zuskov of Voxiva, and Amy Pirretti of ZERO TO THREE for their contributions to developing and implementing the recruitment plan.
Compliance with ethical standards
The recruitment protocol was approved by the George Washington University Institutional Review Board.
Comments to the editor
The findings reported have not been previously published and this manuscript is not being simultaneously submitted elsewhere. Preliminary findings were presented at the Society of Behavioral Medicine’s 37th Annual Meeting in March 2016. The authors have full control of all primary data and agree to allow the journal to review their data if requested.
Funding
Research reported in this publication was supported by the National Institute on Drug Abuse of the National Institutes of Health under Award Number R44DA035017. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Conflicts of interest
Dr. Lorien Abroms has stock options in Voxiva, Inc. and has licensed Quit4Baby and Text2Quit to Voxiva, Inc. Dr. Pamela Johnson and Indira Singh are currently employed by Voxiva, Inc., and Jessica Bushar is employed by ZERO TO THREE, a partner operating the Text4baby service.
Statement of human rights
All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1975 Helsinki Declaration and its later amendments or comparable ethical standards. Animal experimental subjects were not used in this research.
Statement on informed consent
Informed consent was obtained from all individual participants included in the study.
Footnotes
Implications
Practice: An outreach recruitment text message sent to an existing text messaging subscriber list is a novel way to recruit pregnant smokers for a clinical trial.
Policy: Government programs such as Medicaid and other health plans may consider using text messages to identify and reach high-risk populations.
Research: Future research should consider building on the use of text message for recruiting at risk populations for future trials and established treatment programs.
References
- 1.U.S. Department of Health and Human Services. The Health Consequences Of Smoking—50 Years Of Progress: A Report Of The Surgeon General. Atlanta, GA: National Library of Medicine Cataloging in Publication; 2014.
- 2.Tobacco Cessation | Medicaid.gov. Medicaidgov. 2016. Available at: https://www.medicaid.gov/Medicaid-CHIP-Program-Information/By-Topics/Benefits/Tobacco.html. Accessed June 16, 2016.
- 3.U.S. Department of Health and Human Services. Smoking Prevalence and Cessation Before and During Pregnancy: Data from the Birth Certificate, 2014. Hyattsville, MD: National Vital Statistics Reports, Vol. 65, No. 1; 2016. [PubMed]
- 4.Centers for Disease Control and Prevention. Pramstat. Atlanta, GA; 2016: Indicator for whether mother smoked during the last three months of pregnancy.
- 5.Frew PM, Saint-Victor DS, Isaacs MB, et al. Recruitment and retention of pregnant women into clinical research trials: an overview of challenges, facilitators, and best practices. Clinical Infectious Diseases: An Official Publication of the Infectious Diseases Society of America. 2014;59(Suppl 7):S400–S407. doi: 10.1093/cid/ciu726. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Lopez E, Simmons V, Quinn G, Meade C, Chirikos T, Brandon T. Clinical trials and tribulations: lessons learned from recruiting pregnant ex-smokers for relapse prevention. Nicotine & Tobacco Res. 2008;10(1):87–96. doi: 10.1080/14622200701704962. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Cinciripini K, Sorensen G, Klar N, et al. An evaluation of videotaped vignettes for smoking cessation and relapse prevention during pregnancy: the very important pregnant smokers (VIPS) program. Tob Control. 2000;9(90003):61iii–61i63. doi: 10.1136/tc.9.suppl_3.iii61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Emmons K, Sorensen G, Klar N, et al. Healthy baby second-hand smoke study: project brief. Tob Control. 2000;9(90003):58iii–58i60. doi: 10.1136/tc.9.suppl_3.iii58. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Goldenberg R, Klerman L, Windsor R, Whiteside, Jr H. Smoking in pregnancy: final thoughts. Tobacco Control. 2000;9(90003):85iii-86. doi:10.1136/tc.9.suppl_3.iii85. [DOI] [PMC free article] [PubMed]
- 10.Pollack KI, Oncken CA, Lipkus IM, Peterson BL, Swamy GK, Pletch PK, Lyna P, Namenek Brouwer RJ, Fish LJ, Myers ER. Challenges and solutions for recruiting pregnant smokers into a nicotine replacement therapy trial. Nicotine and Tobacco Research. 2006;8(4):547–554. doi: 10.1080/14622200600789882. [DOI] [PubMed] [Google Scholar]
- 11.Ruggiero L, Webster K, Peipert JF, Wood C. Identification and recruitment of low-income pregnant smokers who are we missing? Addict Behav. 2003;28(8):1497–1505. doi: 10.1016/S0306-4603(02)00269-1. [DOI] [PubMed] [Google Scholar]
- 12.Solomon LJ, Secker-Walker RH, Flynn BS, Skelly JM, Capeless EL. Proactive telephone peer support to help pregnant women stop smoking. Tob Control. 2000;9(Suppl III):iii72–iii74. doi: 10.1136/tc.9.suppl_3.iii72. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Strecher VJ, Bishop KR, Bernhardt J, Thorp JM, Cheuvront B, Potts P. Quit for keeps: tailored smoking cessation guides for pregnancy and beyond. Tob Control. 2000;9(Suppl III):iii78–iii79. doi: 10.1136/tc.9.suppl_3.iii78. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Boucher JKonkle A. Understanding inequalities of maternal smoking—bridging the gap with adapted intervention strategies. Int J Environ Res Public Health. 2016;13(3):282. doi: 10.3390/ijerph13030282. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Prenatal Care Utilization, Child Health USA 2013. Mchbhrsagov. 2016. Available at: http://mchb.hrsa.gov/chusa13/health-services-utilization/p/prenatal-care-utilization.html. Accessed June 30, 2016.
- 16.Ingall GCropley M. Exploring the barriers of quitting smoking during pregnancy: a systematic review of qualitative studies. Women and Birth. 2010;23(2):45–52. doi: 10.1016/j.wombi.2009.09.004. [DOI] [PubMed] [Google Scholar]
- 17.Petersen Z, Nilsson M, Everett K, Emmelin M. Possibilities for transparency and trust in the communication between midwives and pregnant women: the case of smoking. Midwifery. 2009;25(4):382–391. doi: 10.1016/j.midw.2007.07.012. [DOI] [PubMed] [Google Scholar]
- 18.Maxwell AE, Bastani R, Glenn BA, Mojica CM, Chang C. An experimental test of the effect of incentives on recruitment of ethnically diverse colorectal cancer cases and their first-degree relatives into a research study. Canc Epidemiol Biomarkers Prev. 2009;18(10):2620–2625. doi: 10.1158/1055-9965.EPI-09-0299. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Hatchett BF, Holmes K, Duran DA, Davis C. African Americans and research participation: the recruitment process. J Black Stud. 2000;30:664–674. doi: 10.1177/002193470003000502. [DOI] [Google Scholar]
- 20.Cleland V, Ball K. Recruiting hard-to-reach populations: lessons from a study of women living in socioeconomically disadvantaged areas of Victoria, Australia. Health Promot J Austr. 2010;21(3):243–244. doi: 10.1071/he10243. [DOI] [PubMed] [Google Scholar]
- 21.Furimsky I, Cheung AH, Dewa CS, Zipursky RB. Strategies to enhance patient recruitment and retention in research involving patients with a first episode of mental illness. Contemp Clin Trials. 2008;29:862–866. doi: 10.1016/j.cct.2008.07.005. [DOI] [PubMed] [Google Scholar]
- 22.Barnett J, Aguilar S, Brittner M, Bonuck K. Recruiting and retaining low-income, multi-ethnic women into randomized controlled trials: successful strategies and staffing. Contemporary Clinical Trials. 2012;33(5):925–932. doi: 10.1016/j.cct.2012.06.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Yancey AK, Ortega AN, Kumanyika SK. Effective recruitment and retention of minority research participants. Annu Rev Public Health. 2006;27:1–28. doi: 10.1146/annurev.publhealth.27.021405.102113. [DOI] [PubMed] [Google Scholar]
- 24.Galli L, Knight R, Robertson S, Hoile E, Oladapo O, Francis D, Free C. Using marketing theory to inform strategies for recruitment: a recruitment optimisation model and the txt2stop experience. Trials. 2014;15:182. doi: 10.1186/1745-6215-15-182. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Turner MM. Using emotional appeals in health messages. In: Cho H, editor. Health Communication Message Design: Theory and Practice. Thousand Oaks: SAGE Publications; 2011. pp. 59–69. [Google Scholar]
- 26.Hoek J, Gifford H, Maubach N, Newcombe R. A qualitative analysis of Maori and Pacific smokers’ views on informed choice and smoking. BMJ Open. 2014;4:e006716. doi: 10.1136/bmjopen-2014-006716. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Nabi R. Exploring the framing effects of emotion: do discrete emotions differentially influence information accessibility, information seeking, and policy preference? Commun Res. 2003;30(2):224–247. doi: 10.1177/0093650202250881. [DOI] [Google Scholar]
- 28.Witte KAllen M. A meta-analysis of fear appeals: implications for effective public health campaigns. Health Educ Behav. 2000;27(5):591–615. doi: 10.1177/109019810002700506. [DOI] [PubMed] [Google Scholar]
- 29.Anderson M. Technology Device Ownership. Pew Research Center. 2015. Available at: http://www.pewinternet.org/files/2015/10/PI_2015-10-29_device-ownership_FINAL.pdf. Accessed June 16, 2016.
- 30.Smith A. Americans and text messaging. Pew Research Center. 2011. Available at: http://www.pewinternet.org/files/old-media//Files/Reports/2011/Americans%20and%20Text%20Messaging.pdf. Accessed June 16, 2016.
- 31.U.S. Department of Health and Human Services. Promoting Maternal And Child Health Through Health Text Messaging. Rockville, MD; 2015:17–26.
- 32.Messages on Critical Issues. Text4babyorg. 2016. Available at: https://text4baby.org/about/message-content. Accessed June 30, 2016.
- 33.Text4baby—Who Involved. Partnerstext4babyorg. 2016. Available at: https://partners.text4baby.org/index.php/about/partners. Accessed June 30, 2016.
- 34.Center for Medicare & Medicaid Services, Maternal and Infant Health. Pilot mobile health program: using mobile technology to reach, educate, and connect pregnant and postpartum medicaid enrollees.; 2016. Available at: https://www.medicaid.gov/medicaid-chip-program-information/by-topics/quality-of-care/downloads/pilot-mobile-health-program.pdf. Accessed June 30, 2016.
- 35.California State University San Marcos. National Survey of Text4baby Participants.; 2013. Available at: http://www.csusm.edu/anthropology/docsandfiles/Text4baby.pdf. Accessed June 18, 2016.
- 36.Abroms L, Johnson P, Heminger C, et al. Quit4baby: results from a pilot test of a mobile smoking cessation program for pregnant women. JMIR mHealth uHealth. 2015;3(1):e1e0. doi: 10.2196/mhealth.3846. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Whittaker R, Matoff-Stepp S, Meehan J, et al. Text4baby: development and implementation of a national text messaging health information service. Am J Public Health. 2012;102(12):2207–2213. doi: 10.2105/AJPH.2012.300736. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Ammerman A, Smith T, Calancie L. Practice-based evidence in public health: improving reach, relevance, and results. Annu Rev Public Health. 2014;35:47–63. doi: 10.1146/annurev-publhealth-032013-182458. [DOI] [PubMed] [Google Scholar]
- 39.Glasgow RE, Vogt TM, Boles SM. Evaluating the public health impact of health promotion interventions: the RE-AIM framework. Am J Public Health. 1999;89:1322–1327. doi: 10.2105/AJPH.89.9.1322. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, Conde JG. Research electronic data capture (REDCap)—a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform. 2009;42(2):377–381. doi: 10.1016/j.jbi.2008.08.010. [DOI] [PMC free article] [PubMed] [Google Scholar]