Background
Recruitment of participants to trials is acknowledged as a challenging facet of research [
1‐
4]. The success of most participatory research and trials primarily depends on effective recruitment and retention of trial participants [
5]. Trials that experience difficulties in recruiting participants have had to extend their recruitment period, vary their recruitment strategy or, in some cases, modify their eligibility criteria in order to achieve adequate sample sizes [
6‐
9].
In the case of pregnant women, recruitment is more challenging to researchers due to the narrow window of eligibility within which to recruit and deliver the intended intervention [
10,
11], time constraints on women with competing priorities [
5,
12,
13], work commitments [
10,
14], responsibilities of caring for other children [
9], and disinterest in research [
9,
15].
To date, a comprehensive process evaluation of the recruitment of pregnant women to a low intensity health promotion trial for infant obesity prevention has not been undertaken. This paper focused on the trial recruitment process of the Communicating Healthy Beginnings Advice by Telephone Randomised Controlled Trial (CHAT RCT) [
16], to identify facilitators and challenges in recruiting pregnant women to a low intensity health promotion programme where health behaviour change messages are delivered via telephone calls or text messages. Recruitment was explored from the perspectives of pregnant women (consenting and non-consenting) and recruiters. The detailed documentation of the recruitment process will assist with translation and replication or for scaling-up of infant obesity prevention trials and interventions to a population level.
Methods
Study setting
The CHAT RCT was conducted across four Local Health Districts within New South Wales in Australia, namely Sydney, South Eastern, South Western and Southern New South Wales Local Health Districts. Pregnant women were recruited at eight hospital sites within the above districts between February 23 and July 27, 2017. The main method of recruitment was opportunistic recruitment at the antenatal clinic waiting rooms by appointed recruiters (three full-time, two part-time and one casual). The CHAT RCT study made use of the opportunity to recruit pregnant women at the time women came to their routine antenatal clinic appointments at the eight hospital antenatal clinics.
Eligibility criteria
Women were eligible to participate in the CHAT RCT if they were 18 years old and over, were in the third trimester of their pregnancy (28–34 weeks gestational age), were able to communicate in English, had a mobile phone and lived in the recruitment area. Women were ineligible if they had a severe medical condition, could not give informed consent, were expecting multiple births or their babies were expected to have major foetal anomalies.
Ethical considerations
The CHAT RCT is registered with the Australian Clinical Trial Registry (ACTRN12616001470482p) on October 21, 2016; Ethics Review Committee of Sydney Local Health District (Protocol No. X16–0360 & LNR/16/RPAH/495).
Study design
Data were collected and analysed from the following three sources: (1) administration of a short questionnaire to women at the time of recruitment to explore women’s reasons for consent and non-consent; (2) interviews with recruiters to capture recruiters’ experiences of the recruitment process; and (3) analysis of field notes taken by the recruiters on the number of women approached and recruited, and the reasons as to why women did not consent to participate.
Phase 1 – Short questionnaire at time of consent
The first data collection phase involved administration of short multiple response and open-ended questions at the time of recruitment to women who consented and to women who did not. Data capture for this phase commenced on June 13, 2017, until the conclusion of trial recruitment on July 27, 2017. The brief questionnaire was administered at the end of the consenting process to consenting women, while non-consenting women were asked if they would please answer the brief questions.
Pregnant women who consented to participate completed three optional questions on Research Electronic Data Capture (REDCap) [
17] using recruiters’ hand-held tablets. REDCap is an electronic data capture tool and a secure web-based application on which the CHAT RCT consent and participants’ data were obtained and captured. Women had the option to choose multiple reasons from a list provided or to provide ‘other’ reasons for participation.
Pregnant women who did not consent to participate were also requested to complete a brief optional questionnaire. Women had the option to choose multiple reasons from a list provided or to provide ‘other’ reasons for non-participation. Data captured on REDCap were exported to Microsoft Office Excel for analysis.
Phase 2 – Interviews with recruiters
Interviews were conducted face-to-face with four recruiters (by ME) during the recruitment phase of the trial to explore their experience of the recruitment process. Recruiters provided written informed consent. Interviews were audio-recorded using a digital recorder and downloaded as voice files, and were between 18 and 30 min in duration. Interviews were semi-structured to explore emerging themes [
18], using an interview guide with seven questions.
To reduce researcher bias, voice files were listened to, coded and analysed independently by two researchers (ME and SM) using the broad research questions as an initial coding framework. Themes were then discussed by the researchers for comparison and consistency of coding, with dominant themes identified and mutually agreed upon. Direct interview quotations illustrate key themes but recruiters’ names were de-identified as recruiters A to D.
Phase 3 – Analysis of field notes
Recruiters wrote hand-written field notes on a purpose-built spreadsheet at point of recruitment. The field notes contained information on recruitment date, recruitment location, number of women who were approached, number of women who consented, number of women who did not consent and recruiters’ perspective of reasons for women not consenting to the trial. The handwritten field notes were later transferred to a combined single Excel spreadsheet. These data were transferred to a Structured Query Language database for analysis of numbers of women approached, numbers of women who did not consent, and recruiters’ perspective of women’s reasons for not consenting. Recruiters’ field notes on women’s reasons for not consenting were grouped into ten broad categories, of which the top four categories were ‘not interested’, ‘not first baby’, ‘limited or no English’ and ‘need more time/information’.
Data collected from the three recruitment phases were triangulated for validation, a strategy that has the benefit of raising the research above narrow interpretations and personal biases that stem from single methodologies [
19].
Discussion
We conducted a process evaluation of recruiting pregnant women to a low intensity health promotion trial for an early childhood obesity prevention trial to identify facilitators and challenges that would contribute to the knowledge base on successful recruitment of pregnant women. The purpose of providing a thorough review of the recruitment process is two-fold, first, for the purpose of replication if translated elsewhere and, second, for scaling-up of the trial to a population level.
Recruitment at antenatal clinics while women waited for their scheduled appointments has proved to be a successful recruitment strategy in other studies [
1,
11,
20]. In this trial, 1155 women were successfully recruited to the CHAT RCT within the stipulated timeframe, which was more than the anticipated number of 1056 women [
16]. The similarity of participants’ characteristics in this study sample to that of the overall CHAT RCT participants suggests that participants’ reasons for consenting to participate are representative of all women who consented to participate in this trial.
Several of the factors that enabled recruitment were similar to those outlined in the literature on recruitment of pregnant women and new mothers to health promotion programmes. Facilitators were the free-of-charge nature of the programme [
21], benefit to women who have had no previous children [
22], altruism [
9,
13,
23], organisational support to conduct the study and to undertake recruitment at antenatal clinics [
9‐
11,
24], and the recruiters’ ability to establish rapport with women [
7,
9,
10,
23,
24]. However, facilitators unique to the success of this recruitment process were the convenience in programme delivery mode via telephone calls or text messages and the nature of the interventions that were not too onerous on women. Converging evidence from all three data collection phases identified that pregnant women considered the mode of intervention delivery via telephone calls or text messages as convenient to them for participation in this trial. Further, pregnant women considered that the trial was not too burdensome and hence were willing to participate. A future challenge is in the subsequent engagement and retention of these women.
Many of the challenges were also similar to those identified in the literature, namely a narrow window of eligibility within which to recruit pregnant women [
10,
11], women’s unwillingness to participate in research [
15,
23], lack of time to complete the survey [
5,
12,
13] and women’s concerns about providing personal information due to privacy issues [
9,
25]. Challenges pertinent to this trial were that women felt they did not need information or support since they had previous children, the trial did not appeal to women from low socioeconomic backgrounds and language barriers of women from non-English speaking background [
9]. Recruiters tried to overcome some of these challenges, but this was not always possible.
Of the 3217 women who were eligible to participate in the CHAT RCT, 1498 women consented to participate in the study, indicating that the CHAT RCT appealed to almost half of women who were eligible. Those women who consented to participate in the CHAT RCT and completed the short purpose-built questionnaire indicated their participation in the study was due to the convenient mode of programme delivery via telephone calls or text messages and suggested that the trial might have value for them.
Only 8% of pregnant women who did not consent to participate in the CHAT RCT completed the questionnaire for non-consenting women and hence their reasons for non-participation might not be representative of all non-consenting women. The low completion rate of the questionnaire (8%) administered to non-consenting pregnant women was similar to that in another study [
15]. However, completion rates may not be comparable to those in many other studies because, in our study, the questionnaire was administered at point of recruitment to explore women’s reasons for not consenting whereas in other studies reasons were explored post-recruitment [
8,
12,
14]. For ethical reasons, the CHAT RCT did not collect demographic or personal information from non-consenters. In order to offset this low response rate, triangulation of these data with the other two data collection phases was undertaken. The inability of some women to converse in English was a major challenge, which prompted further analysis of language spoken by non-consenting women [
9]. Funding constraints precluded employment of sufficient research staff who spoke other languages; implementation strategies should ensure importance of non-English speaking staff, where possible, since language barriers could be more challenging during scaling-up of health promotion interventions into existing services.
Strengths and limitations
A key strength of this research was obtaining the perception of recruitment from those closely associated with the recruitment process (consenting women, non-consenting women and recruiters) and at the time of recruitment. Further, converging evidence through triangulation of the data obtained adds to the strength of this study.
While the mode of intervention delivery via telephone calls or text messages appealed to the vast majority of women, there may be women who are not technologically capable and therefore there is a potential risk of not including them, though we believe this group to be negligible. There was limited participation by those women who did not consent to participate in the CHAT RCT and chose not to complete the short purpose-built questionnaire. These women’s views are not included. There is also a risk of excluding women who do not have the financial capacity to maintain a mobile telephone plan that enables two-way communication with the trial team. Another limitation was the inability of some women to converse in English, which precluded them from participating in this trial.
To address the potential for bias or partiality on the part of the researcher, we used different techniques and methods, including triangulation to cross check findings [
18,
26], and the scrutiny of material by other researchers.
Recommendations
Recruiters believed that the introduction of the study to women by nurses and midwives prior to trial commencement would have allowed women to become familiar with the study and strengthened recruitment efforts [
10,
15,
23]. Cultural translation of booklets to other languages and the delivery of intervention in languages other than English would benefit women from non-English speaking backgrounds and their children.
Conclusion
Facilitators and challenges in recruiting pregnant women to an infant obesity prevention programme were identified through process evaluation of the recruitment phase of the CHAT RCT. Pregnant women demonstrated enthusiasm in participating in the CHAT trial, a low intensity health promotion trial for infant obesity prevention. They were recruited at antenatal clinics while waiting for their scheduled appointments. Despite some challenges in recruiting pregnant women to the CHAT RCT, some of the facilitators in recruitment included mode of delivery of the intervention programme via telephone calls or text messages, the minimal effort required of women to participate, organisational support from the lead site, and recruiters’ knowledge of and commitment towards the trial. Pregnant women’s interest in receiving information via telephone calls or text messages was an indication that women valued the trial.
Acknowledgements
The authors would like to thank the study participants; the Research Assistants for sharing their experience of recruitment; the research team; participating partners of the project from the Sydney, South Eastern Sydney, South Western Sydney and Southern NSW Local Health District; and notably, Sydney Local Health District for organisational and infrastructure support. The CHAT RCT is a partnership project funded by the NSW Ministry of Health.