A Whole New World: Overcoming Methodological Challenges in the Post-Pandemic Recruitment of Pregnant Women at Risk of STIs

Abstract

The impact of the COVID-19 pandemic on recruitment has been significant, including the recruitment of pregnant individuals with behavioral health risk factors. The objective of this manuscript is to highlight our response to the methodological challenges created by the pandemic and how we leveraged technology to adapt and expand our recruitment procedures for the Health Check-up for Expectant Moms, a technology-delivered behavioral skills intervention aimed at empowering and motivating pregnant women engaging in substance use and sexual health risks (e.g., condomless sex, multiple partners) to reduce their likelihood of STIs and substance use during pregnancy. After the onset of the pandemic, we transitioned from in-person recruitment to remote clinic recruitment (e.g., by phone, text messaging, and e-mail) as well as social media campaigns via Facebook^® and Instagram^® with a confidential web-based screening questionnaire for our study targeting pregnant women in 27 counties of Michigan. For remote recruitment, we contacted 6238 women; 1360 completed the study screening questionnaire (mean age = 31.5; SD = 4.6), 90 were eligible, and 73 were enrolled (6% eligibility). For social media recruitment, we had 2512 completing the pre-screening questionnaire. Of these, 501 (mean age = 31.9; SD = 4.6) went on to complete the full study screening questionnaire, 25 women were eligible, and 20 were enrolled (5%). Our 5% eligibility rate from our social media campaign is on par with our remote and in-clinic recruitment methods yet took far less time (15 weeks vs. 150 weeks vs. 48 weeks, respectively). Despite study recruitment challenges related to the pandemic, remote recruitment—particularly social media—was found to be a successful approach, may have a broader reach, and a cost-effective alternative to active recruitment.

The Centers for Disease Control and Prevention (CDC) reported in 2022 that nearly 14% of pregnant individuals reported current alcohol use, and 5% reported binge drinking (four or more drinks on an occasion) in the past month (Gosdin et al., 2022). Cannabis use continues to increase during pregnancy, with 7.9% reporting use (Substance Abuse & Mental Health Service Administration, 2022), and the most common reasons are the relief of pain, nausea or vomiting, and stress or anxiety (Ko et al., 2020). The COVID-19 pandemic led to substantial increases in alcohol and other drug use during pregnancy (Czeisler et al., 2020). Alcohol and drug use are associated with sexual risk-taking, and this cooccurrence can lead to sexually transmitted infections (STIs) among women including during pregnancy (Feaster et al., 2016; Goggins et al., 2020; Rehm et al., 2012). STI risk has been on the rise among childbearing women and the pandemic-related challenges including restrictions to screening and treatment have likely led to substantial increases and continued health challenges among this vulnerable population (Mmeje et al., 2020).

Pregnancy is a unique time to address and prevent the co-occurring risks of substance use and condomless sex, sex with multiple partners, and/or sex with individuals with unknown HIV status, and yet the recruitment of pregnant women for clinical research can present several unique challenges (Cochrane et al., 2022), including social stigma, shame, and guilt. We developed and are currently testing a technology-delivered, low-cost intervention, the Health Check-Up for Expectant Moms (HCEM; see full study protocol) (Tzilos Wernette et al., 2019), to empower and motivate women to engage in behavioral skills (e.g., male and female condom use) to help reduce their likelihood of STIs during pregnancy. The COVID-19 pandemic impacted women’s health research in many ways including our study.

Impact of the COVID-19 Pandemic Across Studies

A number of women’s health trials, including those focused on pregnant populations (Cochrane et al., 2022; Parks et al., 2022; Simo et al., 2022), pivoted in response to the COVID-19 pandemic to implement remote procedures for recruitment, consent, data collection, and intervention activity. This approach has the potential to overcome several barriers, especially as the majority (90%) of those who are of childbearing age report regular use of cell phones and internet-capable devices (Perrin & Atske, 2021). Before the pandemic, Gould and colleagues (2021) approached pregnant women in-person in clinical settings to screen and consent; their modified methods resulted in online pre-screening with recruitment through digital media campaigns including social media platforms. Special consideration was given to protecting the safety of vulnerable populations in recruitment, consenting, and assessment methodology. In a post-pandemic trial of pregnant women in Vancouver, social media was used as a study recruitment tool and was noted to be more efficient and effective compared to more traditional, in-person approaches (Cochrane et al., 2022).

The objective of this manuscript is to highlight our response to the methodological challenges created by the pandemic, and how we leveraged technology to adapt and expand our recruitment procedures for a technology-delivered intervention intended for pregnant women who engage in substance use and sexual health risks (e.g., condomless sex, multiple partners). Furthermore, we compare the screening and enrollment rates and demographic characteristics across our three recruitment approaches (i.e., social media, in-person, and virtual clinic) in our sample of 163 pregnant women.

Methods

Participants

The parent study included a total number of 176 cisgender female pregnant patients between the ages of 18 and 45, reporting ≤ 21 weeks gestation, and meeting specific behavioral risk criteria (including alcohol/drug use and condomless sex, multiple partners). All participants were recruited between April 2019 and September 2023. Before the pandemic (April 2019 through March 2020), study participation was limited to eligible patients within our health system’s primary care clinics providing obstetric care (i.e., obstetricians and family medicine providers). After the pandemic (August 2020 through September 2023), our recruitment strategies were expanded to include women living in one of the partnering counties in the state of Michigan. Recruitment efforts now included engagement via social media (Facebook^® and Instagram^®). Regardless of the recruitment method, all participants were paid $5 to complete the screening questionnaire and up to a total of $220 for completing all the study procedures. During baseline and follow-up assessments, urine samples were collected to diagnose STIs of interest (i.e., Chlamydia trachomatis, Neisseria gonorrhea), and hair samples were collected to corroborate self-reported drug use at follow-up assessments. We received approval from the Institutional Review Board (HUM00143896) before recruitment began and throughout the post-pandemic changes to our study protocol.

Procedure

Initial study recruitment began in April 2019 and was conducted solely in person at our institution’s primary care clinics providing care during pregnancy. With the onset of the pandemic in March 2020, all in-person study recruitment and any participant contact, including new enrollment, were halted per a university mandate. As it was unknown when in-person research activities would resume, we spent approximately five months adapting the study to be completed remotely. Independent of the impact of the pandemic, we recognized that our original screening questions were casting a wide net with regard to sexual health risks (e.g., low threshold for risk). Guided by the literature and consulting with experts in the field, we therefore restricted our screening criteria for sexual health risks to enroll those for whom this intervention was intended (e.g., higher threshold for risk). This included a screening question on partner monogamy, number of male sexual partners within the last six months, and number of new sexual partners.

Because we were no longer meeting participants face-to-face in a private clinic room, we took precautions to help protect the safety and confidentiality of our participants for remote completion. This included using a university-secured and encrypted website (Qualtrics) for the eligibility screener, a HIPAA-compliant communication platform for calls and text messages (Twilio), recruiting only after the second provider visit to ensure an active pregnancy, and implementing new remote protocols for suicidality and child abuse and neglect. We previously detailed how we adapted to the pandemic by modifying our HCEM intervention, conducting study visits remotely, and other changes in our protocol such that we could continue during the pandemic (Tzilos Wernette et al., 2021).

Whether in person or virtual, the recruitment of participants was done using the health system’s electronic health record (EHR). This approach allowed our staff to check clinic schedules and identify new pregnant patients to approach in the clinic (when in-person) or to call. Before the pandemic disruption, research staff would view the weekly clinic schedules through the EHR, searching for new pregnant patient appointments, or patient appointments that were still within our recruitment window (≤ 21 weeks gestation). Our research team made decisions regarding which clinics to visit based on patient volume on a given day. The limitations to this approach include low volume days and the burden on staff for travel time to clinics.

Remote Clinic Recruitment

Remote clinic recruitment began after the onset of the pandemic in August 2020 and was a dramatic change from in-person clinic recruitment. Specifically, potential participants were still identified from clinic schedules, but recruitment took place remotely via telephone calls, text messages, and email messages, allowing us to reach more patients in less time. These additional methods of recruitment communication were not permissible prior to the pandemic and greatly facilitated connecting with potential participants. The procedure included calling patients after their second obstetrics visit to help ensure an established pregnancy. We also adapted our procedures for contact with the participant. For example, if we were unsuccessful at reaching them by phone, we would follow up with a text message and then via email if there was no response to both the phone call and the text messages. We provided the same content in all out-reach attempts, specifically a brief description of the study, including the time commitment and incentive amount. All participants were told that the study aimed to help women have healthy pregnancies and was comprised of an initial 60–90-min visit, two 10-min online booster sessions, and three 30–40-min follow-up visits throughout the pregnancy, earning a total of $220 for their time. Our research staff emailed a copy of the consent form and scheduled a remote baseline visit for eligible patients.

Remote Social Media Recruitment

While remote clinic recruitment had many benefits as compared to in-person recruitment, our team also encountered challenges, including low volume of pregnant patients. To overcome challenges in recruiting from the clinic population, we expanded recruitment beyond our health system, primarily through social media ad campaigns. This expansion required modifications to our study protocol. First, we could no longer collect hair or urine samples at our clinics as we had done previously when pandemic restrictions were lifted. Second, we modified our STI testing from in-person lab testing at our clinic to remote collection with participants receiving at-home STI test kits by mail at the designated time points. These home-based STI test kits were provided by the Johns Hopkins “I Want the Kit (IWTK) project (IWTK).

We launched our social media recruitment in February 2023 and posted ads on Facebook^® and Instagram^® targeting pregnant women ages 18–45. We initially started with the eight counties immediately surrounding our health system and added more distant counties when slowed recruitment necessitated a wider reach. The expansion started with the most populated counties in the state. Furthermore, the addition of each county required extensive research by study staff in order to compile local resources for mental health, sexual health care, and substance use services (per our study protocol). In total, we expanded our social media campaign to include twenty-seven counties within our state. The ads featured pictures of pregnant women from different racial and ethnic backgrounds and contained non-descript taglines such as, “Expecting? Participate in an at-home healthy pregnancy study!” (see Fig. 1).

Fig. 1.

HCEM social media ad example

Interested women clicked on the ad and were brought to a Qualtrics website where they completed a brief pre-screening questionnaire (see Table 1). The pre-screening questionnaire contained a brief introductory paragraph that included study contact information before advancing to the questions. If eligible on the pre-screening questionnaire, women could then leave their contact information for follow-up. The pre-screener contained reCAPTCHA verification and open-text response questions to reduce the chance of the questionnaire being completed by Internet “bots.” Internet protocol (IP) addresses were also checked to ensure pre-screeners were not filled out internationally, and that the same IP address was not completing multiple pre-screeners. Furthermore, by requiring interested women to leave their contact information for a follow-up phone call, the risk of fraudulent responses was significantly reduced. The study team also compiled electronic mail (email) addresses and phone numbers into a reference list in order to check for duplications. The pre-screening questionnaire served two important functions: (1) removing prospective participants who did not meet the gestational age eligibility criteria (≤ 21 weeks), and (2) helping to reduce inappropriate study participation.

Table 1.

HCEM pre-screener questions

Question	Response type
Please click on the box below to continue: “I’m not a robot”	reCAPTCHA verification
Are you currently pregnant?	Yes*/No
Do you live in the state of MI?	Yes*/No
What is your zip code?	Open text response
How many months pregnant are you?	Multiple choice (1–9 mo.) ≤ 5 months*
How many days per week do you do any physical activity? (Please enter number of days from 0 to 7)	Open text response
Have you already participated in the Health Check-up for Expectant Moms Study through Michigan Medicine?	Yes/No*
(If eligible) Great, thank you for your responses! You are eligible to take the study screening survey. This survey will determine if you are eligible to participate in the study or not. Please enter your name and phone number in the space below, so we can call you with next steps. Thank you!	Open text response

Note: The bold* responses denote eligibility

Once contact information was provided, a member of the study team called the potential participant to ask an additional four verification questions: (1) What is your due date? (2) How many weeks pregnant are you? (3) Where are you receiving your prenatal care? (4) Who is providing your prenatal care? Additionally, the study team provided more detailed information about the study. If verified, the candidate would then be emailed the full study screening questionnaire to determine enrollment eligibility (see Fig. 2). To note, there were three individuals who we called and were not able to provide accurate information (e.g., they gave inaccurate responses; falsified attempts). All completed screening questionnaires were paid an incentive as a $5 Amazon^® e-gift card, regardless of study eligibility. Participants were not paid for completing the brief pre-screener. It was only after staff verification that they were offered to take the paid eligibility screener. If eligible, the IP addresses of the completed screener were checked. This was to ensure that the same individual was not attempting to enroll multiple times. Once eligibility was ensured, the prospective study participant was contacted to schedule the baseline study visit, if they had continued interest in participating.

Fig. 2.

Flow diagram of social media recruitment

Results

We report the findings across our three recruitment methodologies (see Table 2). For remote clinic recruitment, we contacted 6238 women, and 1360 completed the study screener (mean age = 31.5; SD = 4.6), 90 were eligible, and 73 enrolled (73/90 = 81%). We had a high social media response rate of 2512 completing the pre-screener. Of these, 501 (mean age = 31.9; SD = 4.6) went on to complete the full study screener, 25 women were eligible, and 20 were enrolled (20/25 = 80%). In comparison to remote recruitment, we approached 860 women during in-person recruitment (pre-COVID), 213 of whom went on to take the screening questionnaire (mean age = 30.4; SD = 5.0). Of these, 126 were eligible, and 70 enrolled (70/126 = 56%).

Table 2.

Comparing three different recruitment methods, procedures, and outcomes

Recruitment method	Dates of enrollment	Setting/modality	Duration of recruitment	Approachedb/ screened c/ eligible/ enrolled	Percent eligible	Percent enrolled
In-person clinica	April 2019–March 2020	Michigan Medicine (MM) clinics; Research staff recruit pregnant patients in clinics	48 weeks	860 / 213 / 126 / 70	59.7%d	60.1%
Remote clinic	August 2020–September 2023	Search Electronic Health Record (EHR); Reaching out directly via phone, email, or text	150 weeks	6238 / 1,360 / 90 / 73	6.60%	81.0%
Social media	February 2023–Septemeber 2023e	Targeted ads on Facebook/Instagram; 27 counties within Michigan	15 weeks	2512 / 501 / 25 / 20	5.00%	80.0%

Notes:

^aPre-COVID pandemic (2019–2020)

^bIdentified by EHR and contacted by staff (either in person or remotely)

^cCompleted full screener

^dEligible prior to restricting eligibility criteria for sexual health risk

^eSocial media recruitment was done intermittently, not in consecutive weeks

Overall, the percentage of women who were eligible on the social media pre-screening questionnaire that went on to take the full screening questionnaire was 45%, almost 20% higher than our in-person recruitment screening rate of 25%. For remote recruitment, this rate (i.e., the number of those screened divided by the number we contacted) was 22%.

We did not find significant differences in age, race, ethnicity, education, or # of weeks currently pregnant between women enrolled via social media (n = 20) versus remote clinic (n = 73) and in-person clinic (n = 70) recruitment (see Table 3). We found that women recruited in person at the clinic were significantly more likely to be either legally married or living together versus women who were recruited remotely or online (p = 0.002).

Table 3.

Demographics characteristics of enrolled participants by recruitment strategy

Characteristics	Recruitment method			P-valuea
	In-person clinic	Remote clinic	Social media
Number of participants N (%)	70 (43)	73 (45)	20 (12)
Age (mean (SD))	30 (4.9)	29 (4.8)	30 (6.8)	0.3547
Race (N (%))				0.2089
Caucasian/White	51 (73)	39 (53)	13 (65)
Black	13 (19)	23 (32)	5 (25)
Other	6 (9)	11 (15)	2 (10)
Ethnicity (N (%))				0.0568
Non-Hispanic	67 (96)	61 (84)	17 (85)
Hispanic/Latina	3 (4)	12 (16)	3 (15)
Current marital status (N (%))				0.0020
Legally married/living together but not married	59 (84)	42 (58)	13 (65)
Separated/divorced/never married	11 (16)	31 (42)	7 (35)
Education (N (%))				0.0665
< 8th grade and HS/GED	8 (11)	20 (27)	3 (15)
Technical/trade school/Some college	14 (20)	16 (22)	7 (35)
College graduate/Postgraduate	48 (69)	37 (51)	10 (50)
# weeks pregnant (mean (SD))	14 (4.0)	13 (3.9)	14 (6.5)	0.4340

^aDescriptive statistics were calculated for demographic variables and compared across three recruitment methods using ANOVA, or chi-square tests, as appropriate

^bDenotes statistical significance

Discussion

In this manuscript, we describe the methodological challenges related to the COVID-19 pandemic that impacted the recruitment of study participants in our ongoing trial of pregnant women. To our knowledge, this is the first study to describe the impact of the pandemic on the recruitment of pregnant women with heightened behavioral risks (e.g., alcohol/drug use and condomless sex, multiple partners), as well as lessons learned to overcome these recruitment challenges.

It became evident in this study that remote/virtual clinic recruitment of pregnant individuals—which was implemented in response to the pandemic—came with a number of challenges. There were declining birth rates both locally and nationally during the pandemic which were attributed in part to reduced employment, and this decline disproportionately affected black women (Hamilton et al., 2023). Given the demographics of our region, these rates affected the number of pregnant individuals through our health system and therefore limited our pool of potential participants. Moreover, we recognize that the disclosure of stigmatizing behaviors can be an obstacle to study recruitment, particularly for pregnant women who struggle with substance use and sexual health risks (e.g., condomless sex, multiple partners, or an unknown HIV status of sexual partners).

Our social media campaign was more effective than our remote clinic recruitment, with respect to the proportion enrolled (80% vs. 55%) and the time to study enrollment after recruitment. Furthermore, our 5% eligibility rate from our social media campaign is on par with our remote and in-clinic recruitment methods, yet took far less time (15 weeks vs. 150 weeks vs. 48 weeks, respectively). These findings mirror prior studies with similar results (Cochrane et al., 2022). In a recent systematic review, social media was found to be significantly more efficient than offline methods for recruitment, and generally more cost effective (Brogger-Mikkelsen et al., 2020).

One challenge with social media in particular is identifying and reducing inappropriate activity or falsified attempts, which is cited as a common problem in trials that use social media for recruitment (Parks et al., 2022). Our study team learned through experience that it was critical to check IP addresses, add a reCAPTCHA tool, include open-text responses, and require interested women to leave their contact information for a follow-up phone call by our staff. Contact information (email, phone) was checked for duplication, and women were asked “verification questions” (e.g., baby due date; how many weeks pregnant; name of provider) which resulted in significant reductions of fraudulent responses. Another important lesson we learned was to restrict incentive payment until after the full study screener had been completed, meaning it was a verified participant. Without these methods in place, numerous screeners were rapidly and fraudulently completed, which threatened to quickly deplete our incentive funds. Similar strategies were used by Parks et al. (2022), including “insider questions,” survey protection against bots (e.g., reCAPTCHA and open-text responses), checking for duplicate IP and email addresses, response consistency checks (e.g., Are the estimated date of delivery and gestational age aligned?), and only providing incentives once eligibility status has been confirmed (Parks et al., 2022).

On balance, a hybrid approach to recruitment offers both advantages (cost-effective, time-efficient) and potential disadvantages. In our study, while there were several verification steps that we had to complete to deter fraudulent activity, the wide reach and rapidity of social media recruitment made it superior to other methods. Remote clinic recruitment involving accessing the electronic health record and clinic schedules also offers a viable approach. A limiting factor of course is the necessary patient volume at study recruitment sites (e.g., hospitals or clinics).

The present study has limitations, including the possibility of selection bias in recruiting our sample online vs. in-person. Remote participation requires access to a phone or device with a reliable internet connection as well as some digital literacy which could exclude under-resourced and marginalized women. We also note that our study team modified our eligibility criteria during the remote clinic recruitment (unrelated to the impact of the pandemic) to restrict our STI risk criteria (and ensure eligible women were at a certain risk threshold). This change may have confounded the significant difference in marital status between in-person vs. remote/online recruitment as we added questions on partner monogamy and the number of recent sexual partners.

The benefits of digital trials are many and have become even more apparent in response to the pandemic (Xue et al., 2020). Digital tools offer an efficient way to recruit samples, and this includes hard-to-reach and diverse populations given the remote nature of their participation (e.g., participants do not need to physically be present at a clinic recruitment site) (Clements & Stem, 2021). Given this relatively nascent field of study, Potočnik and colleagues (2021) outline a summary of recent meta-analyses identifying trends in study recruitment and noted a need for future research to explore the impact of the validity and reliability of technology-driven approaches to recruitment (Potočnik et al., 2021). Our findings add to this literature demonstrating that studies can adapt their protocols to remote recruitment, whether remote clinic or social media, to reliably recruit participants. Remote recruitment has important implications even in our now “post-pandemic” world. Once implemented, these methods can offer efficiency, lessen the burden on study teams and participants, and can also vastly expand the pool of recruits far beyond what is feasible with traditional in-person recruitment methodology.

Importantly, remote recruitment can be generalizable to other vulnerable and marginalized populations (e.g., those struggling with substance use, sexual minorities, and rural or immigrant communities), beyond our study population of pregnant women. Because of the anonymity that online surveys can provide, they may facilitate the disclosure of sensitive information more so than surveys completed in person. This digital nature also allows the participant to choose the space and time that they feel most comfortable for completing surveys and other study tasks, the importance of which cannot be understated. Removing the invisible barrier of perceived stigma and judgement is not only a more person-centered approach, but it could lead to more authentic responses and therefore less bias-prone data. Remote recruitment may also reach non-treatment-seeking populations, or those who have inconsistent or incomplete care.

In this manuscript, we share the ways in which we leveraged technology to adapt and expand our recruitment procedures to overcome barriers created by the COVID-19 pandemic. Our results demonstrate that as compared to in-person clinic recruitment, both remote recruitment and social media (e.g., Facebook^®) recruitment are effective ways to reach and recruit pregnant women who engage in stigmatizing behaviors, including alcohol/drug use and condomless sex, and/or sex with multiple partners. Social media, specifically, was a very efficient and a cost-effective method to reach and recruit study participants. Strategies to increase the detection of inappropriate activity and falsified attempts, which are particular risks for social media, remain important.

Supplementary Material

Supplementary Information The online version contains supplementary material available at https://doi.org/10.1007/s11121-025-01775-1.