Introduction
For nearly two decades, rates of both maternal opioid use-related diagnoses and neonatal opioid withdrawal syndrome (NOWS, formerly known as neonatal abstinence syndrome) have continued to rise [Reference Patrick, Schumacher, Benneyworth, Krans, McAllister and Davis1–Reference Hirai, Ko, Owens, Stocks and Patrick3]. Long-term health and developmental outcomes of children exposed to in utero opioids are unclear, but several studies suggest that children with prenatal opioid exposure remain at increased risk for future adverse health [Reference Hwang, Diop and Liu4–Reference Liu, Kong, Leslie and Corr8], behavioral [Reference Sherman, Ali, Mutter and Larson9–Reference Hall, McAllister and Wexelblatt12], speech-language [Reference Uebel, Wright and Burns10,Reference Hall, McAllister and Wexelblatt12–Reference Fill, Miller and Wilkinson14], and academic outcomes [Reference Fill, Miller and Wilkinson14,Reference Oei, Melhuish and Uebel15]. There is an urgent need for well-designed, prospective, long-term cohort studies to evaluate these findings and the impact of early opioid and other substance exposure on later childhood outcomes.
Understanding the health and neurodevelopmental consequences of in utero substance exposure on children will allow for the development of targeted interventions to improve the long-term well-being and success of these children. Previous work on NOWS outcomes has been criticized for not accounting for the effects of socioenvironmental factors on child health and development, prompting content experts, the National Institutes of Health (NIH), and the Substance Abuse and Mental Health Services Administration (SAMHSA) to call for research evaluating prenatal substance exposure-related outcomes while accounting for highly influential variables [Reference Jones, Kaltenbach, Benjamin, Wachman and O’Grady16–Reference Terplan, Patrick and Re19]. As such, the NIH Helping to End Addiction Long-term® (HEAL) HEALthy Brain and Child Development (HBCD) study aims to apply a multidimensional assessment of perinatal factors that may impact a child’s health and development, including, but not limited to, prenatal substance exposure. It will include collection of medical and family histories; behavioral and developmental evaluations; structural and functional brain assessments (magnetic resonance imaging, and electroencephalography); biospecimen collection; and social and home environment evaluation in order to identify risk and resilience factors that may mitigate adverse outcomes [20]. The HBCD birth cohort study plans to accomplish this aim through the recruitment of a large sample of pregnant and early postpartum persons across the USA and follow-up of the child–caregiver dyads through early childhood.
Recruitment and retention of pregnant individuals and their children into research can be challenging [Reference UyBico, Pavel and Gross21], and prenatal substance exposure further increases the complexity of this research engagement. Individuals with substance use disorders (SUD) often experience socioeconomic difficulties, including unemployment, housing insecurity, transportation barriers, societal stigma, and concurrent mental health issues [Reference Sutter, Gopman and Leeman22–Reference Schiff, Stoltman and Nielsen24], all of which can make involvement in longitudinal research difficult [25]. Additionally, fear of child loss, Child Protective Service involvement, and legal repercussions are specific concerns faced by pregnant persons with SUD [Reference Sutter, Gopman and Leeman22–Reference Schiff, Stoltman and Nielsen24,Reference Faherty, Stein and Terplan26], and make this population particularly vulnerable when approaching them for research involvement. Understanding the additional challenges that may arise in recruiting caregivers with SUD, this study aimed to offset these challenges by summarizing what is already known about factors promoting research engagement among caregivers and their children in longitudinal studies in preparation for the HBCD study [20].
Methods
The authors followed the Preferred Reporting Items for Systematic reviews and Meta-Analyses Extension for Scoping Reviews guidelines [Reference Tricco, Lillie and Zarin27]. The PubMed search engine was used to perform a literature review restricted to articles published from January 1, 2010, through November 23, 2023, in order to focus on modern approaches to participant engagement and because metanalyses and systematic reviews of recruitment and retention published before 2010 are extant in the literature [Reference Booker, Harding and Benzeval28,Reference Teague, Youssef and Macdonald29]. Relevant search terms related to caregivers and/or children were combined first with the term “recruitment” and then “retention” using the following format: (search term) AND (recruitment) or (retention), (Supplementary Table 1).
Eligible articles were published as full text in the English language and described recruitment and/or retention-related findings in longitudinal, observational cohort studies with at least two distinct data collection points involving caregiver–child dyads with children 6 years of age and younger. Titles and abstracts procured using the prespecified search terms were reviewed by five authors (TEC, AJ, JB, CC, and SK) to remove duplicates and ineligible articles. Full-text articles of potentially eligible manuscripts were then reviewed independently for inclusion (Fig. 1). Articles of unclear eligibility were discussed among the authors, and all discrepancies (n = 12) were resolved via consensus.
Preferred reporting items for systematic reviews and meta-analyses (PRISMA) flow diagram.
Data extraction from the included articles was completed by the aforementioned authors and included systematic extraction of information on the study design, goals, participant data collected, population, follow-up duration, percent enrolled and percent retained, and outcomes related to recruitment and retention facilitators and barriers. These data are presented in Table 1 for each included article and summarized across all articles in the text. Given the methodological heterogeneity of the included studies, a formal data pooling and metanalysis were not possible.
Article summaries
a CDGEMM = Celiac Disease Genomic Environmental Microbiome and Metabolomic Study; bTARGet Kids! = The Applied Research Group for Kids Study; cADOS = Autism Diagnostic Observation Schedule-2; dPRIDE = PRegnancy and Infant DEvelopment Study; eNEST = The Newborn Epigenetic STudy; fSES = socioeconomic status; gHLA = human leukocyte antigen; hR&R = recruitment and retention; iDNA = deoxyribonucleic acid.
Results
Article selection
The PubMed database search identified 9,430 articles. After review of titles and removal of duplicates, 2,902 abstracts remained and were reviewed. After application of the exclusion criteria, 237 articles remained for full-text review. Of these manuscripts, 37 met the eligibility criteria and were included in this review (Figure 1).
Characteristics of included studies
Twenty-nine (78%) manuscripts addressed recruitment while 18 (49%) addressed strategies for retention. Thirty-four (92%) articles discussed approaches that support recruitment and/or retention and 18 (49%) addressed factors that hinder recruitment/retention.
The 37 articles included described 34 separate cohort studies from 8 different countries (USA, UK, Australia, Italy, Netherlands, Canada, Germany, and Japan). Results from some studies were featured in several articles (e.g., National Children’s Study [NCS] [Reference Fahrenwald, Wey, Martin and Specker30–Reference Robbins, Bridges, Childers, Harris and McElfish33] and PRegnancy and Infant DEvelopment Study [PRIDE] [Reference van Gelder, Merkus, van Drongelen, Swarts, van de Belt and Roeleveld34,Reference van Gelder, van de Belt, Engelen, Hooijer, Bredie and Roeleveld35]). Methodologies of the studies described within the 37 included articles were heterogeneous in terms of the study goals, length of follow-up, and data collected, and the study sample sizes ranged from 9 to 10,412 participants.
Sixteen articles focused on newborns and their parents, and 12 of those studies approached adults who were pregnant or in the preconception period. Most articles (n = 18) aimed to recruit mothers and their children, while 10 manuscripts were inclusive of either parent and their offspring, and 4 more broadly targeted children and their families. Fewer articles specifically addressed fathers (n = 1) or grandparents (n = 1), and only three articles more broadly focused on legal guardians/caregivers. While all articles included children ≤ 6 years per the review’s eligibility criteria, the age of the children in each study varied, spanning from in utero to 17 years. Three articles addressed underrepresented minorities, and two focused on families from lower socioeconomic backgrounds.
The duration of follow-up ranged from toddlerhood to a goal of 21 years, though one of these prolonged-monitoring studies closed prematurely due to lack of feasibility (National Children’s Study) [Reference Fahrenwald, Wey, Martin and Specker30–Reference Robbins, Bridges, Childers, Harris and McElfish33] and the other is still ongoing, approaching completion of its first decade (PRegnancy and Infant DEvelopment [PRIDE] Study) [Reference van Gelder, Merkus, van Drongelen, Swarts, van de Belt and Roeleveld34,Reference van Gelder, van de Belt, Engelen, Hooijer, Bredie and Roeleveld35]. The number of data collection time points varied from as few as two [Reference Ballieux, Tomalski, Kushnerneko, Johnson, Karmiloff-Smith and Moore36,Reference Bailey, Bann, Bishop, Guarda, Barnum and Roche37] to 46 different family interactions in the aforementioned PRIDE study [Reference van Gelder, Merkus, van Drongelen, Swarts, van de Belt and Roeleveld34,Reference van Gelder, van de Belt, Engelen, Hooijer, Bredie and Roeleveld35]. Types of data collected included questionnaires on demographics, diet, health, education, stressors/conflict, family functioning, environmental exposures, and behavior; biospecimen samples of blood, cord blood, saliva, feces, breastmilk, and cheek swabs; environmental specimens of air, dust, and water; child growth measurements and physical exams; and child evaluations such as assessment of imaging, microbiome, genotyping, motor function, eye-tracking, neuropsychological testing, and developmental performance.
Recruitment and retention considerations
Setting of recruitment
Several factors appeared to influence research recruitment and retention for caregivers and their children. In-person, face-to-face recruitment was found to be very effective [Reference Beasley, Ciciolla and Jespersen38–Reference Menon, Ward and Gaboury41] and may even increase retention and benefit the study long term [Reference van Gelder, van de Belt, Engelen, Hooijer, Bredie and Roeleveld35,Reference Ozonoff, Hill, Hill, Ashley and Young42], but it was both labor-intensive and costly. These factors may lead investigators to consider other less time-consuming approaches [Reference Park, Winglee, Kwan, Andrews and Hudak31,Reference Daly, Levy, Xu, Levy and Fontana39,Reference Bergmann, Keitel-Korndörfer and Herfurth-Majstorovic43], such as mass emailing [Reference Daly, Levy, Xu, Levy and Fontana39] or paid media advertisements like commercials, brochures, or radio advertisements, despite relatively lower enrollment rates with these methods [Reference Beasley, Ciciolla and Jespersen38,Reference Maghera, Kahlke and Lau44]. Other wide-reaching, web-based tools, such as social media platforms [Reference Beasley, Ciciolla and Jespersen38,Reference Gordon, Proschold, Harville, Dutra and Groer45–Reference McGorm, Brown and Thomson48], online advertisements and programs [Reference van Gelder, van de Belt, Engelen, Hooijer, Bredie and Roeleveld35,Reference Gordon, Proschold, Harville, Dutra and Groer45], and web-based questionnaires, [Reference van Gelder, Merkus, van Drongelen, Swarts, van de Belt and Roeleveld34] were also found to be valuable approaches for recruiting caregivers and their children into research studies and supported family convenience for data collection. Postal mailing and telephone calls were generally found to result in lower participant recruitment and were toilsome [Reference Burke-Garcia and Mathew46].
Mailing letters did have its place; however, in raising awareness of a research investigation in a community [Reference Riesch, Ngui and Ehlert32], and community-based recruitment, including posting flyers in public places and in-person recruitment at shopping centers, festivals, parades, and other public events, were effective strategies [Reference Bergmann, Keitel-Korndörfer and Herfurth-Majstorovic43], particularly in rural neighborhoods [Reference Fahrenwald, Wey, Martin and Specker30]. Finally, creating disease-specific registries was useful in increasing knowledge of research investigations and in helping caregivers find available, relevant studies [Reference Greeley, Naylor, Cook, Tucker, Lipton and Philipson49].
Personnel recruiting
Equally important to the setting of participant recruitment was the personnel involved in the recruiting. Medical provider-based sampling achieved high recruitment success [Reference Park, Winglee, Kwan, Andrews and Hudak31]. However, this approach can be burdensome for the clinician and reduce referrals, particularly when the task was perceived as interfering with competing workloads or the clinician–patient relationship [Reference Maghera, Kahlke and Lau44,Reference Wong, Hopkins, O’Donovan, Thapar and Modi50]. There are strategies that could be employed by the study team to minimize this burden. For example, establishing a relationship and workflow with the referring clinicians [Reference Robbins, Bridges, Childers, Harris and McElfish33,Reference Zook, Jordan and Adams51], and seeking referrals without placing the burden of consent on the provider [Reference Maghera, Kahlke and Lau44,Reference Manca, O’Beirne and Lightbody52] were effective in keeping the participating clinician engaged and facilitated recruitment. Other strategies to increase success of enrollment in healthcare settings included designating “a champion” among healthcare personnel to communicate with prospective participants [Reference Giuliani, Bertino and Oberto53] and be solely dedicated to recruitment [Reference Zook, Jordan and Adams51]. Providing reading material in the waiting room followed by a brief discussion and referral by the clinician were likewise effective, capitalizing on patient–clinician relationship and provider involvement [Reference Maghera, Kahlke and Lau44].
Although recruitment directly by the researcher can be efficient, recruitment success can be increased when the research is first introduced to the caregiver by a member of the child/family’s healthcare team [Reference Beasley, Ciciolla and Jespersen38,Reference Menon, Ward and Gaboury41,Reference Woolfenden, Eapen and Axelsson54]. However, this approach may not be universally effective as evidenced by one study that found that using clinical nurses to explain the study followed by referral of interested patients to the research team yielded such low response rates that the research staff turned to approaching caregivers directly [Reference Woolfenden, Eapen and Axelsson54]. Researchers recruiting in the healthcare setting may be especially successful enrolling pregnant persons and their children as the recurrent obstetrical visits afford multiple opportunities for contact [Reference Faro, Sauder and Anderson40].
Other approaches to study enrollment include partnering with nontraditional recruitment personnel. For example, utilizing staff at childcare centers was found to be a useful approach to recruiting children and their guardians into research studies [Reference Ballieux, Tomalski, Kushnerneko, Johnson, Karmiloff-Smith and Moore36]. For caregivers with children who do not participate in formal childcare, playdates at the research lab, where caregivers could socialize with babysitters available to care for the children, were a unique and successful strategy for recruiting young children into investigations [Reference Brand, Gans, Himes and Libster55].
Relationship with study staff
Similar to recruitment, retention of research participants often depended on the caregiver establishing a relationship with the study team [Reference Giuliani, Bertino and Oberto53]. This connection can be fostered by assigning a dedicated study coordinator [Reference Zook, Jordan and Adams51] and building a team of well-trained research staff [Reference Beasley, Ciciolla and Jespersen38]. Additionally, experience of the study team appeared to positively influence participant retention over time, presumably through a more skilled approach and better rapport with study participants [Reference Johnson, Lynch and Baxter56].
Sustaining relationships over time was facilitated by repeated contact with participants between study visits and was shown to improve retention and follow-up [Reference Arora, Manohar and Bedros57], while prolonged time between visits and lost communication was associated with increased study drop-out rates [Reference Bartlett, Kolb and Kingsley58]. This recurring contact could be in person (e.g., accompanying a mother to a Women Infants Children appointment) [Reference Brannon, Kuhl and Boles59], via mail (e.g., holiday or birthday cards for child participants) [Reference Leonard, Kenyon and Valitutti47,Reference Brannon, Kuhl and Boles59], or through communication with the caregiver (e.g., phone, email, or social media connection) [Reference Beasley, Ciciolla and Jespersen38,Reference Leonard, Kenyon and Valitutti47]. Finally, ensuring accurate and up-to-date contact information of the caregiver as well as identifying alternate relations that can be reached and making repeated attempts to connect were noted to be essential for successful retention in longitudinal research [Reference Beasley, Ciciolla and Jespersen38,Reference Zook, Jordan and Adams51,Reference Brannon, Kuhl and Boles59,Reference Sobotka, Lynch, Dholakia, Mayampurath and Pinto60].
Timing of consent
The timing of and approach to consent can be important in optimizing the effectiveness of recruiting caregivers and their children into a study, particularly in a study like HBCD which plans to recruit pregnant and postpartum individuals and those with young children. Attempting study enrollment during the early stages of parenting was found to be overwhelming to some new parents and hindered recruitment rates [Reference Arora, Manohar and Bedros57]. However, minor shifts in approaches, such as obtaining verbal consent in the immediate postpartum period for noninvasive collection of biospecimens in newborns (with the full informed consent and biospecimen linkage to clinical data within the following week), can minimize caregiver burden and be an effective strategy for recruiting newborns and their families [Reference Stroustrup, Bragg and Spear61]. Similarly, poor timing of consent efforts was also found to be a reason for parental disinterest in research participation in scenarios outside the newborn period, specifically when a child was admitted to the hospital and the decision to enroll in a study was time-sensitive [Reference Menon, Ward and Gaboury41]. Allowing caregivers adequate time to make their decision to participate, including time for a parent to discuss the proposal with their partner or other family members, increases the likelihood of study enrollment [Reference Beasley, Ciciolla and Jespersen38,Reference Chudleigh, Hoo and Ahmed62].
Understanding the study
Another aspect that should be considered when approaching enrollment and retention in longitudinal studies is ensuring the caregivers understand the study. Knowledge of a study facilitates recruitment and retention of caregivers and their children through the follow-up period [Reference Bartlett, Kolb and Kingsley58,Reference Chudleigh, Hoo and Ahmed62]. Specifically, comprehending the purpose and importance of the research [Reference Beasley, Ciciolla and Jespersen38,Reference Bartlett, Kolb and Kingsley58], knowing the study procedures and time commitment [Reference Beasley, Ciciolla and Jespersen38], and having access to the study results [Reference Beasley, Ciciolla and Jespersen38,Reference Giuliani, Bertino and Oberto53,Reference Hentschel, Hsiao and Chen63] were noted to be important to individuals approached for recruitment. Additionally, knowing how their child was contributing to a larger goal or “greater good” was a motivator for a caregiver’s ongoing study participation [Reference Giuliani, Bertino and Oberto53], and caregivers were more likely to continue research participation if they recognized the study could potentially help their child or other children [Reference Chudleigh, Hoo and Ahmed62]. Retention was also enhanced when the caregiver felt part of the working alliance and understood their privacy was important to the study team [Reference Sarfi, Eikemo and Konijnenberg64], while apprehension over privacy, confidentiality, and transparency negatively influenced their decision to take part in a study [Reference Beasley, Ciciolla and Jespersen38,Reference Hentschel, Hsiao and Chen63].
Conversely, not fully comprehending the study procedures or the potential benefits of the research was shown to negatively impact recruitment [Reference Menon, Ward and Gaboury41]. Caregivers were also more likely to decline participation if they perceived there was a possible risk to their child by taking part in the study [Reference Beasley, Ciciolla and Jespersen38,Reference Chudleigh, Hoo and Ahmed62], if they were concerned over the invasiveness of biospecimen collection, or if the study would compromise their child’s autonomy [Reference Beasley, Ciciolla and Jespersen38,Reference Hentschel, Hsiao and Chen63].
Minimizing burden
Several studies demonstrated that reducing the burden on families was important in fostering participation in research. Families’ busy schedules, work conflicts, travel distance and transportation issues, lack of childcare, and frequent changes in contact information all served as barriers to recruitment and retention of caregivers and their children into research investigations [Reference Beasley, Ciciolla and Jespersen38,Reference Ozonoff, Hill, Hill, Ashley and Young42,Reference Gordon, Proschold, Harville, Dutra and Groer45,Reference Woolfenden, Eapen and Axelsson54,Reference Brannon, Kuhl and Boles59]. For low-income families in particular, securing transportation, having childcare, and time constraints were found to impact retention [Reference Brannon, Kuhl and Boles59]. To address these logistical barriers, conveniences such as after-hours and weekend calling [Reference Maghera, Kahlke and Lau44,Reference Zook, Jordan and Adams51], study-provided transportation [Reference Beasley, Ciciolla and Jespersen38] and/or covering transportation/parking costs associated with study appointments [Reference Zook, Jordan and Adams51], and providing childcare during data collection improve participation and long-term retention [Reference Beasley, Ciciolla and Jespersen38].
Families can also feel burdened by the data collection process. Study drop out increased when the study was considered too time consuming, questions were perceived as redundant, or the response for data collection was felt to be burdensome [Reference Sobotka, Lynch, Dholakia, Mayampurath and Pinto60,Reference Li, Keown-Stoneman and Borkhoff65]. To address this issue, study teams can utilize programs that were already in place to reduce participant burden caused by data collection [Reference Stroustrup, Bragg and Spear61]. Specific examples include leveraging existing databases and utilizing electronic medical records to extract necessary information (instead of eliciting it from the participants)[Reference Wong, Hopkins, O’Donovan, Thapar and Modi50,Reference Woolfenden, Eapen and Axelsson54,Reference Stroustrup, Bragg and Spear61,Reference Grech, Kizirian and Lal66], completing comprehensive surveys and assessments at the initial visit to decrease the number of follow-up visits for data collection[Reference Stroustrup, Bragg and Spear61], and offering study assessments via phone or other remote options (rather than in-person only) [Reference Thibadeau, Reeder and Andrews67].
Participant retention can also be enhanced by using the conveniences of modern technology for communication and data collection. Short message service (SMS) text messaging and email communication were shown to be a preferred modality of communication for many participants and can be used for greetings, reminders, contact information updates, and even some data collection over the course of a study while being mindful of securing confidentiality and privacy of protected health information [Reference Barry, Sabhlok and Saba68,Reference Nkyekyer, Clifford, Mensah, Wang, Chiu and Wake69]. Digital reminders and multiple SMS text messages were also found to increase clinic attendance [Reference Nkyekyer, Clifford, Mensah, Wang, Chiu and Wake69]. Finally, study websites were proposed as a strategy to keep caregivers up to date and engaged in the research [Reference Arora, Manohar and Bedros57], a strategy proven effective in the general retention literature [Reference Lane, Armin and Gordon70].
Incentives
Incentives and compensation for time and effort are important for participating caregivers [Reference Beasley, Ciciolla and Jespersen38,Reference Zook, Jordan and Adams51,Reference Sobotka, Lynch, Dholakia, Mayampurath and Pinto60], as is the timeliness of the reimbursement [Reference Zook, Jordan and Adams51]. Most commonly, financial compensation (ranging from $20-$75 per session in the reviewed studies) was found to be useful in promoting participation and diminishing the burden of time, lost work, and/or travel required for involvement in the investigation [Reference Beasley, Ciciolla and Jespersen38,Reference Sobotka, Lynch, Dholakia, Mayampurath and Pinto60,Reference Nkyekyer, Clifford, Mensah, Wang, Chiu and Wake69]. Financial compensation was also noted to be useful in encouraging health care providers to refer qualifying patients to the research team [Reference Maghera, Kahlke and Lau44,Reference Manca, O’Beirne and Lightbody52]. Access to free healthcare services was observed to be a research motivator for families from lower socioeconomic backgrounds [Reference Arora, Manohar and Bedros57]. For some individuals, contributing to knowledge in the area of study or helping others acted as an important, nonmonetary recruitment incentive [Reference Chudleigh, Hoo and Ahmed62].
Study population
One of the more difficult aspects of recruitment and retention surround factors that are largely unmodifiable. Maternal lifestyle behaviors, including smoking, alcohol use, and working outside the home during pregnancy, were predictive of study drop-out [Reference Johnson, Lynch and Baxter56]. Similarly, parental demographic variables, such as younger age, unmarried, lower household income, less formal education, un-/under-employed status, and self-identification as a minority, were strong predictors of caregiver–child attrition in longitudinal studies [Reference Johnson, Lynch and Baxter56,Reference Stephenson, Hetherington, Dodd, Mathews and Tough71]. Furthermore, certain populations were found to be less likely to participate in research, such as the socioeconomically disadvantaged, underrepresented minorities, and those individuals for whom a cultural barrier was present [Reference Woolfenden, Eapen and Axelsson54,Reference Williams, Neville, Gillespie, Leary, Hamilton-Shield and Searle72].
Investigators seeking to recruit an inclusive and representative panel of participants will need to be aware of special challenges that may impact an individual who is a minority or holds particular cultural beliefs. Specifically, for some minorities, especially those with limited English proficiency (LEP), lack of study interpreters and availability of printed materials in multiple languages was noted to be an obstacle to successful recruitment [Reference Faro, Sauder and Anderson40,Reference Phillips, Fischer, Baxter, Shafranski, Coe and Kling73]. For populations with differences in cultural customs and beliefs, mistrust of medical experts was found to interfere with participant enrollment [Reference Phillips, Fischer, Baxter, Shafranski, Coe and Kling73]. However, these challenges can be overcome with proper planning. Several studies found that involving community leaders and establishing research partnerships with community organizations serving the population of interest increased minority caregivers’ willingness to participate in research studies [Reference Faro, Sauder and Anderson40,Reference Brannon, Kuhl and Boles59,Reference Phillips, Fischer, Baxter, Shafranski, Coe and Kling73]. For minority caregivers with LEP, having a bilingual researcher or staff who spoke the language of the family or interpreter services was essential in engaging participants and promoting retention [Reference Zook, Jordan and Adams51,Reference Phillips, Fischer, Baxter, Shafranski, Coe and Kling73]. Hiring culturally competent and culturally sensitive staff with strong interpersonal skills to work with families was helpful with retaining caregivers and their children from the minority backgrounds [Reference Zook, Jordan and Adams51].
Investigators need to also consider nontraditional families in which a father or grandparent may have the primary custody of a child. While most studies focused on recruitment of mothers and their children, some specifically addressed engaging fathers or grandparents in research with their children or grandchildren, respectively. For fathers, online crowdsourcing appeared to be a valuable tool in encouraging paternal participation in research investigations [Reference Parent, Forehand, Pomerantz, Peisch and Seehuus74]. While fathers were more apt to drop out early in prospective studies, they were more likely to remain for the duration of the study if they attended the second visit, suggesting that study team efforts should focus on early engagement with and retention of fathers [Reference Parent, Forehand, Pomerantz, Peisch and Seehuus74]. Grandparents were more likely to participate in research if informed about the study by an already-involved parent [Reference Ishikuro, Obara and Osanai75]. Grandparent recruitment also benefited from person-to-person communication (vs email) and the use of different questionnaire modalities (pencil/paper options vs only computer/tablet) [Reference Song and Bonds76].
Discussion
This scoping review synthesizes the data on facilitators and barriers to recruitment and retention of young child–caregiver dyads, providing information that could help improve research engagement of this population in longitudinal cohort studies. The impetus for this review was to prepare for a multicenter prospective, longitudinal study evaluating the complex interplay between innate and extrinsic factors in child development, particularly in children with previous in utero substance exposure, as set forth by the NIH HBCD study [20]. The HBCD-funded cohort of children will be intensely studied and followed from in utero through early childhood. To accomplish this goal, the HBCD study teams will need to successfully recruit and retain a large cohort of pregnant persons, caregivers, families, and their children using the informed approaches set forth in this review.
The methodologic approach was a key factor in the success or failure of recruitment and retention of caregivers and children in the studies included in this review. The HBCD study aims to oversample for children with prenatal substance exposure, and, therefore, the study team will need to deliberately consider how the selected recruitment and retention techniques may apply differently to persons affected by substance use. Results from previous, longitudinal studies on prenatal exposures to marijuana [Reference Fried77,Reference McLemore and Richardson78], cocaine [Reference Graziotti, Hammond and Messinger79], and opioids[Reference Graziotti, Hammond and Messinger79–Reference Hans and Jeremy81] can provides useful guidance on navigating these approaches.
Pregnant and parenting individuals with SUD are a highly stigmatized and vulnerable group[Reference Sutter, Gopman and Leeman22–Reference Schiff, Stoltman and Nielsen24] who have high rates of coexisting mental health disorders [Reference Benningfield, Arria and Kaltenbach82], higher incidences of mental health diagnoses in the first postpartum year [Reference Corr, Schaefer, Hollenbeak and Leslie83], histories of trauma [Reference Ouimette, Kimerling, Shaw and Moos84], and limited social supports [Reference Sutter, Gopman and Leeman22]. Allowing researchers access to their personal lives involves real risks to these individuals with 33 states now having some level of punitive policies in place for substance use during pregnancy despite contrary guidance from professional societies and federal agencies [Reference Faherty, Stein and Terplan26]. It is critical that researchers have a good understanding of each state’s laws and local policies on mandated reporting and differences of such reporting within the research versus clinical settings. Furthermore, longitudinal research can present ethical challenges due to its continuous data collection and its study of vulnerable populations. This concern is particularly relevant in a study that aims to follow persons with both SUD and their children. For individuals affected by SUD, issues may arise related to sustained relationships with research staff, potential for relapse and change in clinical status, and interactions with the legal system [Reference Scott and White85]. Similarly, children participants are considered vulnerable given their inability/limited ability to provide informed consent, their observation over long periods of time, and the potential stress and discomfort of research tasks [Reference Thurman86].
Special precautions need to be taken to ensure the safety of these persons and their children. The HBCD’s broader advisory committees, in collaboration with local community partner-advisors, offer specific recommendations on how to address concerns around the law, the community, relationships, and personal needs and challenges that pertain to individuals with SUD [Reference Goldstein, Nervik and Hagen87]. Additionally, ethically informed research protocols, staff training in monitoring and supervision of participants, and debriefing after critical events aid in protecting individuals with SUD [Reference Scott and White85]. Finally, to safeguard children, researchers should be mindful of infant and child experiences in research studies, and both investigators and participants should conceptualize consent as a continued process in longitudinal research [Reference Thurman86].
There were several specific methodologic strategies identified in this scoping review that can aid investigators in engaging caretakers and their children in research studies. The setting of the recruitment played a key role in the success of study enrollment of caregiver–child dyads. While labor-intensive, face-to-face recruitment largely performed better than more impersonal strategies, such as mass emailing [Reference Daly, Levy, Xu, Levy and Fontana39,Reference Adams, Handley and Heald88]. However, this assertion was not always the case[Reference Bergmann, Keitel-Korndörfer and Herfurth-Majstorovic43] and emphasizes the role of various additional circumstances in a caregiver’s decision to participate in a research study. In every approach, building a relationship with the participant was crucial to successful recruitment and retention. This connection was particularly important for persons with a history of SUD [Reference Sarfi, Eikemo and Konijnenberg64,Reference Goldstein, Nervik and Hagen87].
Though less successful in recruitment, web-based tools have become increasingly popular in research secondary to their ease of use [Reference van Gelder, Merkus, van Drongelen, Swarts, van de Belt and Roeleveld34,Reference van Gelder, van de Belt, Engelen, Hooijer, Bredie and Roeleveld35,Reference Beasley, Ciciolla and Jespersen38,Reference Daly, Levy, Xu, Levy and Fontana39,Reference Gordon, Proschold, Harville, Dutra and Groer45,Reference Burke-Garcia and Mathew46]. Online strategies differ greatly in their uptake by various populations, with social media platforms like Facebook™ achieving recognition for better ability to recruit underrepresented populations [Reference van Gelder, van de Belt, Engelen, Hooijer, Bredie and Roeleveld35], and internet-based approaches performing better at engaging older and more educated individuals [Reference Firestone, Cheng and Pearce89]. Additionally, while online modalities like Facebook™ have effort advantages in early recruitment, participants enrolled through this strategy have poorer retention rates compared to the more time-intensive provider recruitment [Reference van Gelder, van de Belt, Engelen, Hooijer, Bredie and Roeleveld35]. Paid media, including commercials and online advertisement, also performed well, though at an increased monetary expense [Reference Beasley, Ciciolla and Jespersen38,Reference Maghera, Kahlke and Lau44].
While convenient and now widely used, these digital methods require special privacy and confidentiality considerations, particularly for individuals with SUD given the various punitive laws around pregnancy and substance use in several states [Reference Faherty, Stein and Terplan26]. Data from websites and apps can access and store large amounts of data, and protocols that use these modalities need to describe how the website or app functions so that Institutional Review Board (IRB) members and participants may understand the relationship the website/app will have with the research [Reference Buchanan90]. Researchers should contact their IRB during the planning stage as many institutions now have guidelines in place to ensure computer- and internet-based research protocols address fundamental risks (violation of privacy, legal risks, and psychosocial stress) and provide the same level of protection as other research involving human participants [Reference I.R.B.Guideline91].
The personnel in the position of approaching participants for enrollment also played a significant role in the success of recruitment and retention of caregivers and their children. Medical provider recruitment outperformed several other types of personnel in engaging a participant, but was noted to be burdensome for the clinician due to competing workloads and potential encroachment on the provider–patient relationship [Reference Maghera, Kahlke and Lau44,Reference Wong, Hopkins, O’Donovan, Thapar and Modi50]. However, it appears that this barrier could be partially overcome by ensuring the providers were not responsible for providing detailed study information and by offering reimbursement to providers for their time in recruiting patients [Reference Maghera, Kahlke and Lau44,Reference Manca, O’Beirne and Lightbody52]. In contrast to physician recruitment, physician referral after the patient reviewed study-related reading material was found to be effective with less burden on the provider, but it required four times as many attempts at contact by the research team than a referral from a family friend, potentially limiting the utility of this approach by the research team [Reference Maghera, Kahlke and Lau44].
Though the initial introduction of a research investigation by a member of the healthcare team generally produced higher enrollment [Reference Beasley, Ciciolla and Jespersen38,Reference Menon, Ward and Gaboury41,Reference Woolfenden, Eapen and Axelsson54], using clinical nurses to explain the study and refer interested patients to the research team failed to effectively recruit participants, and causing the research staff to contact families directly [Reference Woolfenden, Eapen and Axelsson54]. In this study, nurses felt unprepared to obtain consent directly due to inability to provide sufficient study information and clinical time constraints, an issue that was echoed in other investigations as well [Reference Maghera, Kahlke and Lau44,Reference Wong, Hopkins, O’Donovan, Thapar and Modi50]. Response rates may have been improved by providing a family with a decision aid [Reference Bailey, Bann, Bishop, Guarda, Barnum and Roche37], designating a nurse whose sole role was communication with the participants [Reference Giuliani, Bertino and Oberto53], or, if feasible, assigning staff exclusively dedicated to recruiting patients [Reference Zook, Jordan and Adams51].
Timing of approach also plays an important part in the success of caregiver–child recruitment into research studies. Efforts at enrolling during the initial phases of child-raising was often noted to be difficult as caregivers were preoccupied by more immediate issues of caring for a newborn [Reference Arora, Manohar and Bedros57]. When interviewed, caregivers indicated that other times, such as during pregnancy or after the first several weeks of the baby’s life, might be more effective in recruiting families [Reference Arora, Manohar and Bedros57]. Additionally, when a child is admitted to the hospital, as would be the case for infants being observed or treated for opioid withdrawal, parental stress was found to negatively influence a caregiver’s willingness to participate in a research investigation [Reference Menon, Ward and Gaboury41]. In both these scenarios, a potential solution may be to consider initial verbal consent followed by formal, informed consent at a later, less-stressful time [Reference Stroustrup, Bragg and Spear61].
The burden placed on the caregiver and child during the actual execution of the study can be overlooked and addressing these issues proactively can increase the successful recruitment and retention of caregivers and their children into longitudinal research studies. Families’ busy schedules, work conflicts, transportation issues, and lack of childcare were frequent barriers to caregiver-child recruitment and retention [Reference Beasley, Ciciolla and Jespersen38,Reference Ozonoff, Hill, Hill, Ashley and Young42,Reference Gordon, Proschold, Harville, Dutra and Groer45,Reference Woolfenden, Eapen and Axelsson54,Reference Brannon, Kuhl and Boles59]. These challenges are particularly pertinent for persons with SUD as they often experience socioeconomic difficulties such as unemployment, housing insecurity, and transportation barriers [Reference Sutter, Gopman and Leeman22,25]. Similarly, for low-income families these issues were especially challenging as were lack of reliable phone service and the inability to provide consistent contact information due to unstable life circumstances [Reference Brannon, Kuhl and Boles59].
To minimize these inconveniences for caregivers and their children and encourage enrollment and retention, several approaches have been noted to be helpful. Specific solutions include incorporating evening/after-hours and weekend calling [Reference Maghera, Kahlke and Lau44,Reference Zook, Jordan and Adams51], offering participation through phone calls when feasible [Reference Thibadeau, Reeder and Andrews67], allowing flexibility in scheduling or who brings the child participant [Reference Beasley, Ciciolla and Jespersen38,Reference Sobotka, Lynch, Dholakia, Mayampurath and Pinto60], providing childcare during data collection [Reference Beasley, Ciciolla and Jespersen38], and covering transportation and parking costs[Reference Zook, Jordan and Adams51] or providing transportation [Reference Beasley, Ciciolla and Jespersen38]. For families struggling with reliable communication, successful strategies to promote study retention include obtaining alternative contacts and updating them regularly[Reference Zook, Jordan and Adams51] and maintaining regular contact in the follow-up period to avoid possible attrition [Reference Beasley, Ciciolla and Jespersen38,Reference Giuliani, Bertino and Oberto53,Reference Arora, Manohar and Bedros57,Reference Brannon, Kuhl and Boles59,Reference MacBean, Drysdale, Zivanovic, Peacock and Greenough92]. Finally, incentives and compensation for time and effort can offset the burden of the study protocol [Reference Beasley, Ciciolla and Jespersen38,Reference Zook, Jordan and Adams51,Reference Sobotka, Lynch, Dholakia, Mayampurath and Pinto60]. However, special attention must be paid to the issue of coercion, especially in disadvantaged and vulnerable populations such as those with SUD. Disparities in income, paid time off, childcare, and transportation challenges make it difficult to determine a universal incentive for participant time and effort [Reference Langer, Castro and Chen93]. Helpful solutions to this challenge include seeking input from community advisory boards and considering a sliding scale for compensation so that lower income populations receive higher compensation in an effort to overcome barriers to participation that may not be present for those with higher incomes, more stable homes, and access to more resources [Reference Williams and Walter94].
Several studies pertaining to recruitment and retention of caregiver–child dyads specifically focused on minority families, and inclusion of underrepresented minority populations in research is a priority for the NIH as enrollment barriers are particularly prevalent in minority communities [95,Reference Brewster, Steinberg and Magnani96]. For some minority families, cultural and linguistic barriers were found to negatively affect enrollment and study retention [Reference Faro, Sauder and Anderson40,Reference Arora, Manohar and Bedros57]. These challenges can be overcome but require dedicated efforts from the research team in securing bilingual staff and interpreter services[Reference Phillips, Fischer, Baxter, Shafranski, Coe and Kling73] and garnering support from community gatekeepers in order to engage the residents of their community [Reference Faro, Sauder and Anderson40,Reference Brannon, Kuhl and Boles59,Reference Phillips, Fischer, Baxter, Shafranski, Coe and Kling73]. Further, peer navigators have also been shown to increase engagement of vulnerable, hard-to-reach populations [Reference Zgierska, Gramly and Prestayko97,Reference Zgierska, Hilliard, Deegan, Turnquist and Goldstein98].
There are several limitations to this literature review. First, the approach for the review was not a systematic review or meta-analysis, and as such, can be prone to author bias and interpretation. However, the review was undertaken in a standardized fashion, with clear inclusion and exclusion criteria, and any uncertainty over an article’s eligibility for inclusion was reconciled by discussion among the study team. In an effort to identify factors that would be helpful in the development of a longitudinal cohort starting in early childhood or infancy, the inclusion criteria mandated that the article reviewed involved children 6 years of age or younger. Therefore, the facilitators and barriers to recruitment and retention identified in this review may not be broadly applicable to engaging caregivers and their older children in research. Additionally, the scoping review utilized PubMed database; articles indexed outside PubMed may have been missed. Despite these limitations, this study entailed a structured and purposeful review and synthesizes the literature on recruitment and retention of caregivers and their children into longitudinal research. Consistent with a recent systematic review on recruitment and retention strategies of pregnant persons [Reference Goldstein, Bakhireva and Nervik99], this study emphasizes that utilization of a variety of recruitment and retention methods is more likely to achieve a large, population-representative cohort [Reference Maghera, Kahlke and Lau44,Reference Williams, Neville, Gillespie, Leary, Hamilton-Shield and Searle72], The information from this review will be useful in future, prospective cohort studies aiming to study the specific population of caregivers and their children.
Conclusion
Numerous factors affect engagement of caregivers and their children in observational cohort studies. Our findings indicate there is no clear, single strategy that is universally effective in engaging caregivers and their children, and multiple approaches should be considered and tailored to the specific populations of interest in the HBCD study. The information here may inform future study designs aimed at the prospective analysis of long-term outcomes in caregivers and their children.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/cts.2024.624.
Acknowledgments
The authors have no specific acknowledgements.
Author contributions
Dr Tammy Corr is responsible for the conception and design of this review, collection, and critical interpretation of the literature reviewed, analysis of the results, drafting the manuscript, revising and finalizing the manuscript for submission, and takes responsibility for the manuscript as a whole. Drs. Alma Jusufagic, Catherine Caldwell, Steven King, and James Basting are responsible for the collection and critical interpretation of the literature reviewed, analysis of the results, and critically reviewing the manuscript. Dr Aleksandra Zgierska is responsible for the conception of this review, critically reviewing the manuscript, and overseeing the study. All authors reviewed and approved the final manuscript for submission.
Funding statement
Support for this work was provided by the funding from the National Institutes of Health (U01DA055361) and from the National Center for Advancing Translational Sciences grant (U54 TR002014-05A1). The content, views, and opinions presented in this manuscript are those of the authors only and do not necessarily represent the views, official policy, or position of the NIH or any of its affiliated institutions or agencies.
Competing interests
The authors declare none.
Open access
