Lessons learned on recruitment and retention in hard-to-reach families in a phase III randomised controlled trial of preparatory information for children undergoing general anaesthesia

Recruitment and retention are often perceived as challenges to conducting clinical trials [1‐4]. Many trials are unable to recruit sufficiently to meet the requirements of the a priori power calculation, and close prematurely or require an extension [3‐6]. Several reviews including a Cochrane systematic review, state that little is known about successful recruitment to RCTs [2, 6‐9]. Recruitment to paediatric trials in particular, has a history of low participation and faces specific challenges including the ethical issues of recruiting children [7, 10‐14]. More paediatric clinical trials are required to ensure that children and families are receiving acceptable care and gaining access to new opportunities in health [8], so it is important that strategies for successful recruitment and retention are examined. The difficulties of recruiting and retaining families can be further heightened by the challenges that disadvantaged families face [15‐19]. These families - often termed ‘hard to reach’ - are less likely to access services, despite needing them more than others and they would particularly benefit from improved interventions [18‐25]. They are also less likely to participate in studies due to both logistical and attitudinal barriers [16‐18, 20, 26]. It is likely that without the inclusion of such families, the generalisability of any findings is limited [17, 19, 24, 25].

A separate but related issue concerns retention of participants within a trial. Retention rates in clinical trials tend to be poorly reported [26] but previous studies have indicated that attrition can be as much as 70%, which is likely to have a significant impact on the generaliasability of any study findings [24, 26, 27].

Previous research indicates that strategies to facilitate recruitment and retention include: staff working out of hours; face to face recruiting; recruiter(s) developing a relationship with recruitment site staff, recruiters having strong interpersonal skills, good engagement with each participant equally to help them learn about the health issue being addressed; ensuring little or no extra effort is required on the part of participants; conducting a pilot study, and discussing the advantages and disadvantages of participation [4, 9, 24, 28‐31]. Financial incentives have previously been reported to improve participation [6, 18, 19, 29, 32], but this is not common practice and raises the ethical issue of the possibility of coercion [4].

The purpose of this paper is to present the lessons learned from recruitment and retention of families to a clinical trial of preparatory information for children undergoing general anaesthesia (GA) for tooth extraction [33].

The Phase III RCT [33] recruited families whose five to seven year old children were undergoing GA because a child in the family needed teeth extracted due to serious decay. A recent UK audit of GA service provision for dental surgery has shown the national yearly dental case load total of 111,600, of which 60% were children [34]. For full details of the trial methodology see Hosey et al. [33]. Families with a child eligible to enter the trial were referred by primary care dentists to the hospital based specialist paediatric dentist service. The aim of the RCT was to understand whether a preparatory computer package accessed from home and based on cognitive behavioural principles influenced children’s ability to cope with the GA induction and various other elements of the experience. Parents were asked questions regarding their children’s overall behaviour. Children were asked about their anxiety levels and their feelings about going to the dentist and then observed throughout their time in the hospital, particularly the child’s behaviour before, during and after anaesthetic induction. These families were then asked about the hospital experience, whether the preparatory information assisted the families, and about physical and psychological morbidity following the procedure. There were four data collection time points:

Initially, it was planned that this study would take 18 months to complete recruitment and, despite some obstacles (including cancelled theatre lists that were essential to collecting trial data), the recruitment target was obtained in 17 months. Recruitment was conducted in a small waiting room outside a nurse led medical pre-assessment clinic for children who were having GA. This occurred 1 day a week from 7.45 until 17.30 within the hospital day surgery unit. Our target children, those whose dental health was so severe as to necessitate having teeth extracted under dental GA, made up between half and three-quarters of all families who attended each clinic. Children attending the service, both locally and nationally, are typically 6 years of age and have an average of seven primary teeth removed. The dental diagnosis of this level of tooth decay is known as Early Childhood Caries. The Association of Paediatric Anaesthetists of Great Britain and Ireland Guidelines [35] have recommended that these children have access to preoperative preparation. When considering recruitment and retention in this trial, it is important to acknowledge the connection between early childhood caries and socio-economic deprivation [25, 36‐40]. The families referred for this treatment have high levels of caries, and the catchment area includes some of the worst areas of social deprivation in London. Indeed, more than a third of children in the catchment area would be considered to be living in deprivation [37].

Five lessons were learned for promoting recruitment, as follows:

As with all clinical trials, retention proved challenging. In this RCT, participants were asked to participate in a telephone interview both 48 h and 1 week after their child’s GA visit. This was to assess physical morbidity, psychological morbidity and family satisfaction with the dental general anaesthesia treatment and also with the preparatory information that was provided. During the pilot study, parents were asked to choose the time they would prefer to be telephoned and most responded that “any time” was suitable. However, the response rate at 48 h was 56% and, at 1 week only 58.5% successfully completed the telephone call interview. Although most studies deem 50–80%, is acceptable [53, 54], the current RCT had hoped for better. The Centre for Evidence Based Medicine at Oxford University uses 20% attrition to define trial quality [55].

Therefore, the goal for this RCT was a 80% or higher success rate at follow-up. As such, two key strategies were implemented:

These two strategies improved response rates considerably: data were collected at the 48 h telephone call by 151 of the 176 participants (a 85.5% success rate) and the 1 week telephone call had a 83.1% success rate (146 participants contributed data). Using telephone calls for data collection following a hospital event, has the advantage of enabling multiple attempts to make contact and well as giving the family flexibility in timing the delivery of their response. Some studies have used home visits to achieve this and have also reported high retention rates. One such study reported a 87.5% retention rate at the 48 h home visit [56] while another reported a 96.1% retention rate at a 48 h home visit [57]. Although arrangements for home visits are likely to be more difficult, particularly in socially deprived neighbourhoods, they do seem to result in high rates of retention. Perhaps this ‘personal’ approach is highly valued by these families? Conversely, other studies have suggested that offering the follow up telephone calls rather than a hospital or home visit may have indirectly facilitate recruitment because it did not cause inconvenience [9].

Surprisingly, a non-response at the telephone follow-up interview was not related to any demographic variable. In particular: Chi-square analyses exploring the relationship between response to the follow up interview and demographic characteristics revealed that having a mother at home (all participants lived with their mothers) or father at home (p = 0.24); and it was not related to the number of children in the family (p = 0.42). Moreover, the number of teeth that were extracted did not influence a lack of response, either at home (p = 0.40) or immediately post-operatively (p = 0.10). The parental education level also had no influence (p = 0.93).

The following variables were found to enhance retention rates:

The excellent recruitment and retention rates found in this phase III RCT were the result of strategies that were identified both from the previous literature and developed in the pilot study. The participating families are ‘hard-to-reach’ and were drawn from areas of socio-economic deprivation and have been traditionally characterised as having poor engagement with clinical services, let alone research. Underpinning many of the strategies identified in this study is the importance of positive communication and in the relationship between the researchers and the participants [1, 24]. However, there is a potential for conflict here, in that whilst a good relationship might increase recruitment and retention, it might also increase the likelihood of a Hawthorne effect. Many of these families might value the personal approach that participation in a study such as this will offer.

Future studies might report on how much each of the aforementioned strategies contribute to families’ ongoing participation and retention by embedding studies of recruitment strategies within the larger clinical trial.