Research | Open | Open Peer Review | Published:
Participant recruitment and retention in longitudinal preconception randomized trials: lessons learnt from the Calcium And Pre-eclampsia (CAP) trial
Trialsvolume 18, Article number: 500 (2017)
The preconception period has the potential to influence pregnancy outcomes and randomized controlled trials (RCTs) are needed to evaluate a variety of potentially beneficial preconception interventions. However, RCTs commencing before pregnancy have significant participant recruitment and retention challenges. The Calcium And Pre-eclampsia trial (CAP trial) is a World Health Organization multi-country RCT of calcium supplementation commenced before pregnancy to prevent recurrent pre-eclampsia in which non-pregnant participants are recruited and followed up until childbirth. This sub-study explores recruitment methods and preconception retention of participants of the CAP trial to inform future trials.
Recruiters at the study sites in Argentina, South Africa and Zimbabwe completed post-recruitment phase questionnaires on recruitment methods used. Qualitative data from these questionnaires and quantitative data on pre-pregnancy trial visit attendance and pregnancy rates up to September 2016 are reported in this paper. RStudio (Version 0.99.903 https://www.rstudio.org) statistical software was used for summary statistics.
Between July 2011 and 8 September 2016, 1354 women with previous pre-eclampsia were recruited. Recruitment took 2 years longer than expected and was facilitated mainly through medical record/register and maternity ward/clinic-based strategies. Recruiters highlighted difficulties associated with inadequate medical records, redundant patient contact details, and follow-up of temporarily ineligible women as some of the challenges faced. Whilst the attendance rates at pre-pregnancy visits were high (78% or more), visits often occurred later than scheduled. Forty-five percent of participants became pregnant (614/1354), 33.5% (454/1354) within 1 year of randomization.
In preconception trials, both retrospective and prospective methods are useful for recruiting eligible women with certain conditions. However, these are time-consuming in low-resource settings with suboptimal medical records and other challenges. Trial planners should ensure that trial budgets cover sufficient on-site researchers with pre-trial training, and should consider using mobile phone and web-based electronic tools to optimize recruitment and retention. This should lead to greater efficiency and shorter trial durations.
Pan-African Clinical Trials Registry, Registration Number: PACTR201105000267371. The trial was registered on 6 December 2016.
Preconception is now recognized as a period with the potential to influence pregnancy outcomes and long-term child health . Whilst certain interventions recommended during pregnancy might also have an impact if commenced before pregnancy, evidence on preconception interventions is scarce, particularly from low- and middle-income countries (LMICs). Preconception interventions that might influence pregnancy outcomes include interventions targeting chronic conditions such as epilepsy, hypertension, and diabetes , nutritional and/or lifestyle interventions for underweight and overweight populations , interventions to reduce substance use , thromboprophylaxis to reduce pregnancy losses [5, 6], interventions targeting infections such as HIV [7, 8], and interventions to reduce pre-eclampsia [3, 9]. There is, therefore, a need for well-conducted, preconception randomized controlled trials (RCTs) to improve preconception guidance, and various preconception trials are underway [10,11,12].
However, RCTs are known to have difficulties associated with recruitment, compliance, and retention of participants over extended study periods, which can sometimes lead to early trial closures. In addition, participant attrition can cause methodological problems that influence the study results . Preconception RCTs are arguably the most challenging type of RCT to conduct, as eligible non-pregnant women need to be recruited and retained until conception occurs (if and when), and then further, throughout pregnancy, to delivery or beyond. A variable proportion of randomized (non-pregnant) women, therefore, are not included in the final sample of pregnant women, and time to conception cannot be estimated – unlike time to delivery. Recruitment for preconception RCTs is particularly difficult because so many pregnancies are unplanned and highly eligible women may only access the healthcare system once they are already pregnant. Thus, when planning a preconception RCT it is important to consider strategies to optimize participant recruitment and retention.
We designed a RCT known as the Calcium And Pre-eclampsia (CAP) trial to test the hypothesis that calcium supplementation commenced before pregnancy will reduce the incidence of recurrent pre-eclampsia more effectively than supplementation starting at 20 weeks’ gestation . As with preconception folic acid supplementation [14,15,16,17,18,19], if preconception calcium supplementation is shown to be effective, it could have important implications for food fortification policies, particularly in countries with low dietary calcium intake . This article describes a sub-study of the multi-country CAP trial, exploring the recruitment methods and participant retention in the preconception phase of this trial including difficulties and challenges experienced.
The aim of this sub-study was to explore recruitment methods and retention of participants in the CAP trial to inform future preconception trials conducted in LMICs.
Trial design and participants
The World Health Organization CAP trial is a multi-centre, double-blind, parallel-arm, placebo-controlled randomized trial of long-term calcium supplementation in women at high risk of pre-eclampsia. Non-pregnant women were eligible if they had developed pre-eclampsia in their most recent pregnancy, were planning to become pregnant and were willing to provide written informed consent.
This multi-country trial includes one site in Zimbabwe (comprising Harare and Mbuya Nehanda Maternity Hospitals), four sites in South Africa (Frere and Cecilia Makiwane Hospitals in East London, Chris Hani Baragwanath Hospital in Johannesburg, Tygerberg Hospital near Stellenbosch, and Mowbray Maternity Hospital in Cape Town), and a site in Argentina (Institute for Clinical Effectiveness and Health Policy) comprising four hospitals (Hospital Italiano, Hospital San Justo, and the Center of Medical Education and Clinical Investigations in Buenos Aires, and the Institute of Maternity and Gynecology, Nuestra Señora de Mercedes in Tucumán province). Most of these referral hospitals serve mainly low-income populations with low dietary calcium intake.
Participants in the study group received 500 mg of elemental calcium daily (in the form of calcium carbonate) from randomization (preconception) until 20 weeks’ gestation, whereas participants in the control group received identical-looking placebos. All women received unblinded calcium supplementation (1.5 g elemental calcium daily) from 20 weeks’ gestation until delivery as per WHO recommendations for prevention of pre-eclampsia .
The primary outcome of this trial is pre-eclampsia; secondary outcomes include pregnancy, miscarriage, maternal and neonatal complications related to pre-eclampsia, and compliance. For a complete list please refer to the published protocol .
The sample size calculation was informed by a previous WHO study of calcium supplementation from 20 weeks’ gestation in which the incidence of hypertension (with or without proteinuria) among relatively low-risk pregnant women in South Africa was 14% . Women with pre-eclampsia in a preceding pregnancy have a very high risk of recurrence, approaching 50% in some studies . Therefore, for the power calculation, we assumed the incidence of pre-eclampsia in our trial, which involves only high-risk women, to be 25%. To show a reduction in pre-eclampsia to 15%, we calculated that we needed 540 participants with pregnancies continuing beyond 20 weeks’ gestation using Epi Info™ software (CDC) (alpha = 5%, beta = 80%). We anticipated that 50% of women recruited would become pregnant during the study. Therefore, allowing for a miscarriage rate of 15% and loss to follow-up of 10%, we calculated that we needed a sample size of approximately 1440 non-pregnant women.
The methods for this double-blind RCT have been described . Various recruitment methods were proposed in the protocol, and study sites customized the methods to their individual settings. Non-pregnant women attending screening and subsequent research clinic visits were offered compensation at each visit for travel expenses.
To facilitate screening and recruitment, the screening form grouped eligibility criteria into two sections according to whether a woman was permanently ineligible (e.g., did not have previous pre-eclampsia) or temporarily ineligible (e.g., not in a sexual relationship). Women in the latter group could be invited for another screening visit at a later stage. Following randomization, participants were required to attend the research clinic visits every 12 weeks for follow-up from preconception through to delivery. Between-visit contact was to be maintained by 4-weekly telephone calls.
Case Report Form data were entered and validated in an online data management system (OpenClinica; www.openclinica.com) by researchers at the sites.
For this exploratory, mixed-methods sub-study, researchers responsible for recruitment at each site were asked in July 2016 to complete a questionnaire on the recruitment methods used and the “pros” and “cons” of each method. Qualitative data compiled from these questionnaires were tabulated. RStudio (Version 0.99.903 https://www.rstudio.org) statistical software was used for summary statistics. For retention calculations using data up to September 2016, the denominator excluded women who became pregnant. As the randomization code has not yet been broken, no comparative data were analyzed.
Recruitment commenced in July 2011 and was completed on 8 September 2016, taking 2 years longer than anticipated and involving more sites than originally planned.
Table 1 shows the findings of the recruiter-completed questionnaire, highlighting the challenges encountered by different recruitment approaches, lessons learnt, and practical considerations for future trials. Most participants were recruited either by searching past medical records and maternity registers to identify and make contact with potentially eligible women, or by visiting hospital postnatal wards and clinics to approach women who had recently experienced pre-eclampsia. The main advantage of the former approach was the potentially immediate access to large databases of eligible women; the main disadvantage was the huge recruiter workload involved in identifying, pre-screening, and making contact, and the low response and recruitment rates following contact. Medical records and registers often lacked accurate contact details and diagnoses and this contributed to the low recruitment rates with this method. In addition, it required high levels of coordination with local hospital staff not involved in the trial to facilitate access to medical records. As this comprised additional work for hospital staff, without additional compensation, it was not easy for recruiters to implement and sustain.
Prospective visits to wards and clinics were also very time-consuming. Eligible recruits were fewer than with the retrospective method, with many women being temporarily ineligible. However, yields were higher, probably due to the personal, face-to-face approach, which facilitated better communication and time for discussion with those women who were interested. Women recruited with this method represented potential “future” rather than immediate recruits, and time needed to be invested with regular follow-up telephone calls during their “ineligible” phase. This probably contributed to the longer-than-expected trial duration.
In total, 2563 women were screened and 1354 eligible non-pregnant women were recruited and randomized. Participant characteristics stratified by country can be found in Table 2. The median time since participants’ last pre-eclampsia-complicated pregnancy was 10.5 months (interquartile range (IQR) 29.4). Among participants recruited from African sites, poor previous pregnancy outcomes were common with more than half of African participants having experienced a fetal death, and more than a quarter having experienced eclampsia and/or HELLP (hemolysis, elevated liver enzymes and low platelets) syndrome, in their last pregnancy.
Excluding the short first and last recruitment years, the average annual recruitment over the 4-year period from 2012 to 2015 was 298 participants per year (range 218 to 369). Overall, approximately one out of every two women screened was recruited; however, the proportion of women recruited out of those screened increased over time (Table 3).
Approximately 78% (961/1231) of randomized participants attended their first pre-pregnancy visit (PPV), theoretically 12 weeks after randomization, and participant attendance was above 86% for the second PPV (709/821) and all subsequent PPVs. However, timely attendance (within 2 weeks of the scheduled date) was lower, ranging between 57% (698/1231 at PPV 1) and 83% (24/29 at PPV 14), with higher compliance observed over time.
By September 2016, 45.3% of participants (614/1354) had become pregnant with 33.5% (454/1354) of participants becoming pregnant within 1 year of randomization, representing 73.9% (454/614) of all pregnancies. The total number of woman-years of follow-up to achieve one pregnancy was 3.7 woman-years, and to achieve a valid outcome (defined as delivery of a live child or miscarriage or stillbirth with gestational age of 20 weeks or more) was 4.8 woman-years.
To our knowledge, this is the first study of recruitment and retention strategies for preconception trials in LMICs. A range of recruitment approaches was used in this trial that built upon our experiences and lessons learnt from previous RCTs conducted among pregnant women [21, 23]. The majority of participants were recruited following identification from medical records/registers or from ward/clinic-based strategies, which enabled the accrual of a target sample of highly eligible women. However, the slower-than-anticipated accrual and the difficulties experienced by recruiters highlights that many LMIC trial procedures can be improved upon, in particular, by the deployment of more trial staff dedicated to specific recruitment and retention activities and the use of modern technology to facilitate these activities.
A report of a United States trial on preconception low-dose aspirin that used provider/clinic and community-based outreach recruitment approaches concluded that the former was most successful and cost-effective strategy . Unfortunately, in our trial the source of recruits was not prospectively recorded; therefore, it was not possible to correlate recruitment approach with effectiveness. However, it is evident from the qualitative data that certain tasks, particularly the process of identifying and contacting women through hospital records and databases, were very laborious. One of the challenges for recruiters was the high proportion of redundant contact details and inaccurate recorded diagnoses leading to the inability of recruiters to contact many potentially eligible participants and to their incorrect identification and eligibility status. Retrospectively searching records to 5 years was probably too far back to search in our settings and searching a shorter period might have yielded similar participant numbers with a reduction in time invested. Few participants were recruited by self-referral in response to advertising or community outreach strategies. Researchers at one South African site reported a good response following promotion of the trial on a radio talk-show, and this and/or radio advertising may be a worthwhile option to consider in future trials, particularly as radio is an enduring and ubiquitous media in developing countries . Women who refer themselves might also be more motivated to comply with trial visits than those recruited by other methods.
As anticipated, attendance at PPVs was not ideal and recruiters struggled to keep track of telephone contacts and visit schedules for both potential and randomized participants. Such difficulties occurred partly as a result of over-enthusiastic recruitment initially, whereby some women who probably were not particularly interested in the trial were recruited. Over time, however, recruitment and retention improved as recruiters learnt to identify those women genuinely interested in participating, and to pace recruitment efforts against existing follow-up visits. This evolved approach probably accounts for the improved recruitment rate over time. Concerning retention, one South African site reported that mobile phone applications (e.g., Whatsapp) were valuable for following up participants who missed visits as, with such applications, it was possible to see whether a telephone number was current and if the participant had read the message; it was also less time-consuming than a telephone call. In future, to better facilitate recruitment and retention processes, customization of appropriate mobile phone and web-based electronic tools (calendars, diaries) for eligibility screening, follow-up of temporarily ineligible women, and for scheduling follow-up visits and telephone calls with study participants would be a worthwhile investment in LMIC settings, as well as pre-trial recruiter training.
The high proportion of participants recruited through maternity wards and clinics probably contributed to the prolonged accrual time. Most postnatal women at the African sites are offered long-acting hormonal contraception before discharge and such women would have been identified as temporarily ineligible and followed up at 3-monthly intervals thereafter. Routine pre-pregnancy counseling as practiced in certain countries, such as China, helps to facilitate participant recruitment in preconception trials , and recruitment would no doubt have been easier if the CAP trial sites offered routine pre-pregnancy counseling for women with previous pre-eclampsia. However, even with this strategy, a certain proportion of eligible women would be missed because so many pregnancies in our settings are unplanned.
True to our sample size calculation, about half of the recruited women fell pregnant. As most pregnancies (73.9%) occurred within 12 months of randomization, investigators may wish to consider the cost-effectiveness of following up non-pregnant women beyond this or another time-point. In addition, when the planned stopping date is in sight, e.g., 6 months away, it would be prudent to have a strategy in place to stop follow-up of non-pregnant women to avoid investing unnecessary effort.
This sub-study highlights that the most important resources for effective recruitment and retention in preconception trials are motivated and trained human resources. Searching and screening eligible participants who are not yet identifiable by the occurrence of a pregnancy is a very laborious process. However, with a combination of retrospective and prospective approaches, it is possible to yield a sample of highly eligible non-pregnant women. Investigators and funders of future preconception trials in low-resource settings should budget for sufficient on-site researchers and pre-trial researcher training to optimize recruitment and retention. In addition, they should consider using mobile phone and web-based electronic tools. Such deployment should lead to greater recruitment and retention efficiency and shorter trial durations.
Calcium And Pre-eclampsia
Hemolysis, elevated liver enzymes and low platelets
Low- and middle-income country
Randomized controlled trial
World Health Organization
Mumford SL, Michels KA, Salaria N, Valanzasca P, Belizán JM. Preconception care: it’s never too early. Reprod Health. 2014;11:73.
Lassi ZS, Imam AM, Dean SV, Bhutta ZA. Preconception care: screening and management of chronic disease and promoting psychological health. Reprod Health. 2014;11 Suppl 3:S5.
Dean SV, Lassi ZS, Imam AM, Bhutta ZA. Preconception care: nutritional risks and interventions. Reprod Health. 2014;11 Suppl 3:S3.
Lassi ZS, Imam AM, Dean SV, Bhutta ZA. Preconception care: caffeine, smoking, alcohol, drugs and other environmental chemical/radiation exposure. Reprod Health. 2014;11 Suppl 3:S6.
Ismail AM, Hamed AH, Saso S, Abu-Elhasan AM, Abu-Elghar MM, Abdelmeged AN. Randomized controlled study of pre-conception thromboprophylaxis among patients with recurrent spontaneous abortion related to antiphospholipid syndrome. Int J Gynecol Obstet. 2016;132(2):219–23.
Schisterman EF, Silver RM, Lesher LL, Faraggi D, Wactawski-Wende J, Townsend JM, et al. Preconception low-dose aspirin and pregnancy outcomes: results from the EAGeR randomised trial. Lancet. 2014;384(9937):29–36.
Lassi ZS, Imam AM, Dean SV, Bhutta ZA. Preconception care: preventing and treating infections. Reprod Health. 2014;11 Suppl 3:S4.
Mugo NR, Hong T, Celum C, Donnell D, Bukusi EA, John-Stewart G, et al. Pregnancy incidence and outcomes among women receiving preexposure prophylaxis for HIV prevention: a randomized clinical trial. JAMA. 2014;312(4):362–71.
Hofmeyr GJ, Lawrie TA, Atallah AN, Duley L, Torloni MR. Calcium supplementation during pregnancy for preventing hypertensive disorders and related problems. Cochrane Database Syst Rev. 2014;6:CD001059.
Jiang H, Xiong X, Su Y, Zhang Y, Wu H, Jiang Z, et al. A randomized controlled trial of pre-conception treatment for periodontal disease to improve periodontal status during pregnancy and birth outcomes. BMC Pregnancy Childbirth. 2013;13:228.
Skau JK, Nordin AB, Cheah JC, Ali R, Zainal R, Aris T, et al. A complex behavioural change intervention to reduce the risk of diabetes and prediabetes in the pre-conception period in Malaysia: study protocol for a randomised controlled trial. Trials. 2016;17(1):215.
Hofmeyr GJ, Novikova N, Singata M, Fawcus S, Oyebajo A, Munjanja S, Belizán JM. Protocol 11PRT/4028: long term calcium supplementation in women at high risk of pre-eclampsia: a randomised, placebo-controlled trial (PACTR201105000267371). Lancet. http://www.thelancet.com/protocol-reviews/11PRT-4028. Accessed 15 Dec 2016.
Hindmarch P, Hawkins A, McColl E, Hayes M, Majsak-Newman G, Ablewhite J, et al. Recruitment and retention strategies and the examination of attrition bias in a randomised controlled trial in children’s centres serving families in disadvantaged areas of England. Trials. 2015;16:79.
De-Regil LM, Peña-Rosas JP, Fernández-Gaxiola AC, Rayco-Solon P. Effects and safety of periconceptional oral folate supplementation for preventing birth defects. Cochrane Database Syst Rev. 2015;12:CD007950.
Centers for Disease Control and Prevention. Folic acid. https://www.cdc.gov/ncbddd/folicacid/recommendations.html. Accessed 20 Sept 2016.
World Health Organization. WHO Guideline: Daily iron supplementation in adult women and adolescent girls. Geneva: World Health Organization; 2016. http://www.who.int/nutrition/publications/micronutrients/guidelines/daily_iron_supp_womenandgirls/en/. Accessed 20 Sept 2016.
Wilson RD, Audibert F, Brock JA, Carroll J, Cartier L, Gagnon A, et al. Pre-conception folic acid and multivitamin supplementation for the primary and secondary prevention of neural tube defects and other folic acid-sensitive congenital anomalies. JOGC. 2015;37(6):534–52.
Cawley S, Mullaney L, McKeating A, Farren M, McCartney D, Turner MJ. A review of European guidelines on periconceptional folic acid supplementation. Eur J Clin Nutr. 2016;70(2):143–54.
Food Fortification Initiative. http://www.ffinetwork.org/global_progress/. Accessed 20 Sept 2016.
World Health Organization. WHO recommendations for prevention and treatment of pre-eclampsia and eclampsia. Geneva: World Health Organization; 2011. http://www.who.int/reproductivehealth/publications/maternal_perinatal_health/9789241548335/en/, Accessed 15 Sept 2016.
Hofmeyr GJ, Mlokoti Z, Nikodem VC, Mangesi L, Ferreira S, Singata M, et al. Calcium supplementation during pregnancy for preventing hypertensive disorders is not associated with changes in platelet count, urate, and urinary protein: a randomized control trial. Hypertens Pregnancy. 2008;27(3):299–304.
Dildy III GA, Belfort MA, Smulian JC. Pre-eclampsia recurrence and prevention. Semin Perinatol. 2007;31:135–41.
Villar J, Abdel-Aleem H, Merialdi M, Mathai M, Ali MM, Zavaleta N, et al. World Health Organization randomized trial of calcium supplementation among low calcium intake pregnant women. Am J Obstet Gynecol. 2006;194:639–49.
Lesher LL, Matyas RA, Sjaarda LA, Newman SL, Silver RM, Galai N, et al. Recruitment for longitudinal, randomised pregnancy trials initiated preconception: lessons from the effects of aspirin in gestation and reproduction trial. Paediatr Perinat Epidemiol. 2015;29(2):162–7.
UNESCO. Statistics on radio. 2013. http://www.unesco.org/new/en/unesco/events/prizes-and-celebrations/celebrations/international-days/world-radio-day-2013/statistics-on-radio/. Accessed 20 Sept 2016.
We thank the CAP trial researchers at the sites: Emilia Makaza, Eunice Tahuringana, and Bothwell Guzha (Harare, Zimbabwe); Catherine Parker and Gift Phoramphai (Johannesburg, South Africa); Patience Moloi, Annemarie Greef, and Saadiqa Alie (Cape Town, South Africa); Xoliswa Williams, Pamela Njikelana, and Catherine Parker (East London, South Africa); Erika von Papendorp (Stellenbosch, South Africa); Rossana Chalha, Gabriela Castro, and Sebastián Diaz (Tucumán, Argentina); Ricardo López, Paula Rubinstein, Nicole Minckas, Gustavo Izbizky, Javier Schwartzman, Hugo Krupitzki, Juliana Botbol, and Silvana Martin (Buenos Aires, Argentina). We especially thank all the women in Argentina, South Africa, and Zimbabwe who took part in the trial.
The CAP trial is part of the PRE-EMPT (pre-Eclampsia – Eclampsia, Monitoring, Prevention and Treatment) study and is supported by the University of British Columbia, a grantee of the Bill & Melinda Gates Foundation (Global Health Grant Number OPP1017337). UNDP, UNFPA, UNICEF, WHO, World Bank Special Programme of Research, Development, and Research Training in Human Reproduction, Department of Reproductive Health and Research supported the involvement of WHO staff in this study. Support was also received from the Argentinian Fund for Horizontal Cooperation of the Argentinian Ministry of Foreign Affairs, South-South and Triangular Cooperation (FOAR) (no grant number) and the Research Council of Norway, through its Centers of Excellence scheme and the University of Bergen (UiB), Norway, and the Centre for Intervention Science in Maternal and Child Health (CISMAC; project number 223269).
Availability of data and materials
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.
Ethics approval and consent to participate
Ethical approval for the trial was obtained from the Human Research Ethics Committee of the University of the Witwatersrand, Johannesburg (certificate number M10977), the Faculty of Health Sciences Human Research Ethics Committee, Cape Town (HREC 457/2010), Health Research Ethics Committee 1 (certificate number M12/01/002), the Medical Research Council of Zimbabwe (certificate number MRCZ/A/1599), and the Argentinean National Ethical Committee, Centro de Educación Médica e Investigaciones Clínicas (CEMIC), (Protocol Number: 725). In addition, the trial protocol was approved by the Research Project Review Panel of the UNDP/UNFPA/UNICEF/WHO/World Bank Special Programme of Research, Development, and Research Training in Human Reproduction at the Department of Reproductive Health and Research of WHO and the WHO Research Ethics Review Committee, Geneva, Switzerland (certificate number A65750). All participants enrolled in the trial provided written informed consent.
Consent for publication
The authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.