The Food, Feelings, and Family Study: comparison of the efficacy of traditional methods, social media, and broadcast email to recruit pregnant women to an observational, longitudinal nutrition study

Abstract

Background: It is well known that recruitment is a challenging aspect of any study involving human subjects. This challenge is exacerbated when the population sought is reticent to participate in research as is the case with pregnant women and individuals with depression. This paper compares recruitment methods used for the Food, Feelings, and Family Study, an observational, longitudinal pilot study concerning how diet and bisphenol A exposure affect maternal mood and cognitive function during and after pregnancy.

Methods: Pregnant women were recruited to this study over a period of 15 months using traditional methods, social media including paid and unpaid posts, and emails broadcast to the university community. Contingency analysis using the Pearson's Chi-square test was used to determine if recruitment method was associated with likelihood of participation. T-tests were used to analyze Facebook advertisement success. ANOVAs and Fisher exact tests were used to determine if recruitment method was related to continuous and categorical demographics, respectively.

Results: Social media resulted in the largest number of recruits, followed by traditional methods and broadcast email. Women recruited through social media were less likely to participate. In contrast, use of broadcast email resulted in a smaller pool of recruits but these recruits were more likely to be eligible for and complete the study. Most women recruited via social media were the result of unpaid posts to the study's Facebook page. Paid posts lasting at least 4 days were the most successful. Recruitment method was not associated with participant demographics.

Conclusions: Social media has the potential to recruit a large pool of potential subjects; however, when studies require a large time investment such as the case here, women recruited through social media are less likely to participate and complete the study than women recruited through other means.

Trial registration: N/A. This study does not describe a health care intervention.

Keywords: Depression; Nutrition; Pregnant; Recruitment; Social media.

Fig. 1

Study timeline versus number of screening surveys completed. Black, white, and gray bars indicate the number of participants that completed the screening survey as a result of traditional recruitment practices, social media, and broadcast email, respectively, as described in the Methods section. Social media refers to both Facebook and Instagram. Alterations to the study protocol to increase participation are indicated by arrows. Letters on the x-axis indicate the first letter of each month of the year, starting with April represented by the letter A

Fig. 2

Sample recruitment flyer and example of a Facebook paid promotion. The recruitment flyer (a) contains the Food, Feelings, and Family logo and contact information (university logo is not shown due to copyright) and information regarding the study’s objective, participation expectations, incentives, eligibility criteria, and investigator information. The Facebook promotion (b) displays the study’s logo and objective, a link to the screening survey, contact phone number, and text indicating where emojis depicting fruits and vegetables were located. All images other than the university logo (a) and the emojis (b) were original or in the public domain

Fig. 3

Flowchart of participation by recruitment methods from screening to study completion. The term “Social Media” includes both Facebook and Instagram. Diagonal arrows indicate the number of participants that changed from in person to online interviews and vice versa. Percentages represent the percent of total recruits and may not add up to 100 due to rounding

Fig. 4

Geographical distribution of recruits and participants in the Food, Feelings, and Family Study. Google maps was used to indicate locations across (a) the United States, (b) Texas, and (c) Central Texas. Red circles indicate women that completed the screening survey (recruits). Green circles display the location of study participants. In (C) blue triangles indicate the location of cities and towns in Central Texas. The three recruits that completed the screening survey and lived outside the US are not included. The map of the United States was modified from the original [47]. The maps of Texas and Central Texas were derived from [48]. These maps are available under the GNU General Public License