Recruitment and retention: How can we do better?

pharmafile | July 16, 2018 | Feature | Business Services, Manufacturing and Production, Medical Communications, Research and Development, Sales and Marketing |  clinical trials, feature, patient recruitment, patients, pharma 

In pharma R&D, data is king, but consistent struggles with patient retention is undermining the integrity of clinical findings. Matt Fellows speaks to experts from the University of Liverpool to explore what can be done to reverse the trend.

Research and development has always dwelt at the heart of the pharmaceutical and life sciences industries, and that crucial process hinges upon the robust execution of clinical trials to deliver valuable and meaningful data. Only with thorough, representative data sets can clinical development teams build medicines to the highest standards of efficacy and safety. And one of the most – if not the most – important commodity in this system is the patient; only from them can this data be gathered.

But it isn’t as simple as just amassing a sizeable group of candidates; the group must be large enough in size to be statistically significant, it must be representative of a diverse range of biological profiles, and then the participants must be retained at various key checkpoints along the span of a trial in order to create an accurate picture of a drug’s efficacy and safety over a considerable duration of use. Without this, attrition occurs and the integrity and utility of a trial and its findings can be jeopardised. Pharmafocus spoke to Professor Carrol Gamble, Professor of Medical Statistics/Co-Director of the Clinical Trials Research Centre and Anna Kearney, Research Assistant, both from the University of Liverpool’s Department of Biostatistics, to better understand the impact that failures in recruitment and retention of trial participants can cause:

“If a trial does not include sufficient participants in its analysis then this reduces statistical power, meaning that the trial is less likely to detect a difference between treatments if one exists,” Professor Gamble explained. “However, it may also impact the generalisability of results, and in the case of retention, introduce bias if there is a greater loss of participant data in one trial arm, or if the lost data is associated with a particular outcome.

“In the case of low recruitment, one common solution is to extend the recruitment period, but this increases the cost and resources needed for the trial and the time required to answer the clinical question. In the worst cases poor recruitment can lead to early termination of the study.”

Kearney adds: “Sample size calculations are often adjusted upwards to account for expected levels of missing data, placing an added burden on recruitment.”

As damaging as this situation can be to the veracity of the data a clinical trial generates, it is not an uncommon problem, as Kearney notes:

“The extent of recruitment challenges within the UK were highlighted in the mid-2000s. Less than a third of trials recruiting between 1994 and 2002 achieved their original recruitment target with many seeking additional funds or grant extensions. More recent analysis showed a slight improvement with 55% of trials achieving their targets and less trials revising their sample size. 

“This problem is not limited to the UK, with 26% of publicly funded trials in Switzerland prematurely discontinued due to slow recruitment,” she added.

“Retention levels are much harder to assess in published reports, making it difficult to identify the extent of the problem. CONSORT guidelines require the reporting of loss to follow up and exclusions. However, the complexities of reporting data availability at multiple time points or for different analysis populations alongside the numbers who discontinued the intervention can lead to incomplete reporting of missing data.” 

Professor Gamble also notes: “Further blurring of retention is compounded when imputation is used without clarity on the number of participants for whom the results were actually observed or measured.”

“Analysis of published trials estimate that on average 10% of data is missing from trials,” Kearney continues. “Whilst this may not be a cause for alarm, there can be great variability for different populations and health areas. For example, mental health studies, obesity studies and those involving complex interventions are known to be more susceptible to attrition.”

Identifying strategies today

The prevalence and severity of the issue is a cause for concern, and this is compounded by the fact that we often have comparatively little understanding of the multitude of reasons why these failures in retention occur. With such a glaring challenge facing the industry and leading to a considerable wastage of resources and the infirmity of data which can result, there is a real need for practical, tangible strategies in order to address the problem.

“The lack of evidence for effective retention strategies is a major challenge for improving retention rates,” Kearney remarks. “Current practice is often guided by Investigator’s personal experience, trial and error and informal feedback from trial sites from previous trials.”

It was for this reason that Gamble, Kearney et al conducted their study Identifying research priorities for effective retention strategies in clinical trials. The investigation honed its focus on the UK clinical trial landscape to identify prevalent retention methodologies currently in practice, and evaluate where the industry’s priorities lie in examining the effectiveness of such practices.

“We were aware that trials were using ad hoc strategies to try and address retention challenges. These are rarely described in trial reports, and so we wanted to gain a more comprehensive understanding of what retention strategies were being used across the UK,” explained Kearney. “The study involved a survey of NIHR HTA-funded Chief Investigators and UK Clinical Research Collaborative Registered Clinical Trial Units (CTUs). The latter was a more substantial survey to determine how frequently retention strategies were used and whether there was any evidence for their effectiveness.

“Following this we undertook a Delphi survey to gain consensus across UKCRC CTUs as to which retention strategies should be prioritised for evaluation. Retention strategies have resource implications and so it is important to know that these investments are worthwhile. Studies evaluating strategies up to that point had predominantly focussed on increasing questionnaire return despite a great range of causes of attrition within clinical trials.”

Illuminating obstacles

These endeavours helped to shine a light on some of the biggest challenges faced by clinical trial investigators when it comes to fostering continued cooperation and participation with patients throughout the course of a study, and even some of the problems inherent in attempting to evaluate the effectiveness of current strategies to tackle these challenges.

“Randomised studies of retention interventions need to be nested within host clinical trials to give firm evidence as to whether strategies are effective. The number of Studies Within A Trial (SWATs) are increasing and recently the NIHR HTA programme announced a pilot funding scheme to support this research,” Kearney expounded. “However, interventions need to be nested in large trials or across multiple smaller trials in order to have sufficient power for analysis. Investigators may also be reluctant to add in SWATs to avoid increasing burden at sites or introducing complexity to the host trial.

“In addition, many investigators commented to us how there is no blanket solution for retention. The choice of outcome, the timing and method of measuring the outcome, the patient population, health area and nature of the intervention all affect how susceptible a trial is to attrition. The causes of attrition are also broad, including failure to retain participants through either formal withdrawal or loss of contact; patients not returning data; patients failing to attend visits where data measurements are taking place; clinical staff forgetting to collect data within time windows; and technology or laboratory failures. Consequently, a large number of strategies need evaluating across a range of host trial designs.  A retention strategy that is effective in one trial may be ineffective for other designs or patient populations, making building an evidence base particularly challenging.

“Methodological research of recruitment and retention has increased, creating a growing body of published literature. However, identifying solutions within this literature is time consuming and not always practical for investigators. We have developed an online searchable database of recruitment literature (www.orrca.org.uk) to help investigators identify trial specific recruitment strategies and understand the evidence base for their use. This database will soon be extended to include retention literature.”

The study revealed a number of common strategies employed by the industry with the goal of improving retention of their clinical trial participants, including newsletters to sites, a timeline of the participant visit schedule, telephone reminders, use of a pre-paid envelopes, and data collection scheduled with routine care and site training. Professor Gamble noted: “The use of routinely collected data may lead to improvements, but problems exist about completeness and quality of this data, and timely access which may limit widespread use.”

Kearney added: “Chief Investigators mentioned the importance of close monitoring to address retention issues as they emerge, as well as maintaining good relationships with research sites.”

It’s all in the design

The study also found that 78% of respondents reported that they kept the challenges of retention squarely in mind from the earliest stages of a trial, implementing strategies into the design of a study from the outset. On the issue of leveraging solutions from the initial design phases, Kearney remarked: “Most trials experience missing data in some form and so retention strategies must absolutely be embedded into the design of a clinical trial. FDA and EMA guidance strongly emphasise the need to mitigate missing data in trial design and conduct because statistical techniques can never accurately account for it. Recruitment and retention need to be considered at all stages of trial design and then closely monitored throughout trial conduct in order to adapt trial procedures where needed.

“During the early stages of trial design, careful consideration needs to be given to core components such as the choice of outcomes, timing and method of data collection, as well as the participant demographic. Feasibility work is essential to test proposed recruitment and data collection procedures and identify any potential treatment preferences that might affect post-randomisation withdrawals.

“Our paper lists a wide range of retention strategies that can be embedded within trials to try and address specific retention concerns,” she continues. “Strategies can be aimed directly at participants: maintaining regular contact with participants outside of data collection time points; taking multiple contact details; offering flexibility around appointment times and visit locations; incentives; and including prepaid envelopes with questionnaires. Or they can be aimed at the research sites: clear communication of retention procedures during trial start-up training; regular newsletters; a timeline of participant visit/data collection time points; monitoring visits; and investigator meetings.

“Research has shown that there is a correlation between retention rates and the number of strategies employed. When choosing strategies, equal consideration needs to be given to retaining participants and maximising data collection to ensure all relevant data from retained participants is available and eligible for analysis.”

What else can help?

There are any number of reasons why a participant may drop out of a continuing trial. Where a common requirement for them to visit clinical centres in order to follow up with trial organisers could be a key factor in their failure to continue contact, could the convenience of accessible mobile technology provide a solution? It’s a tool continues to offer new ways to tackle age-old problems in healthcare now that smartphones are so ubiquitous and app design continues to impress, so could they help too in the pursuit of improved retention rates?

“Mobile technology/apps theoretically minimise patient burden and could be a valuable retention strategy given the rise of wearable technology and the widespread use of smartphones,” Kearney mulls. “However, patient data return is particularly susceptible to attrition and so whether these methods can live up to their potential remains to be seen. Data protection, reliability of technology, demographic use, integrity of the data and cost effectiveness need to be explored.”

What about increased regulation or standardisation when it comes to bolstering retention? While it may sound appealing on the surface to provide blanket guidance to trial investigators across the board, the reality may be a little too diverse and unpredictable for such a catch-all solution, Kearney suggests.

“Retention challenges are complex. Solutions need to be tailored to individual trials to address potential causes of attrition known to exist for the patient population, health area, intervention, measurement of the outcome, making it difficult to introduce industry wide standards,” she says. “Standards may be of value in ensuring sufficient thought is given to retention problems within trial development. For example, ensuring feasibility work and patient involvement within the design stages are embedded to review proposed procedures and identify recruitment and retention problems early on.

“Data protection regulations including the recently implemented General Data Protection Regulation (GDPR) are impacting retention,” she continues. “Maintaining positive relationships with participants is vitally important, but CTUs often avoid collecting participant contact details, meaning repeated attempts to collect data falls to research sites. Further guidance on trial protocol development is needed to ensure retention procedures are adequately documented. For example, interviews with research staff have highlighted uncertainties of how many times patients should be contacted to try and collect data.  In the absence of information within the protocol, practice was often directed by personal opinion; consequently, approaches varied massively. Experienced research nurses often felt more comfortable pursuing data on multiple occasions, whilst others reported only contacting participants once to collect data for fear of upsetting them. 

“Existing standards need further development to improve reporting of levels of attrition within individual trials,” she added. 

Keep pressing forward

There is clearly more work to be done, but the hard work from Gamble, Kearney et al helps to paint an invaluable portrait of current UK practice with regard to recruitment and retention in clinical trials, and suggests that similar investigations into the prevailing practices of other regions such as the US could be beneficial, with room to take the research to a deeper level in future.

Particularly, it could be suggested that more evaluation may be necessary into the challenges and effectiveness of the strategies uncovered in the study to determine which produce better results than others – or perhaps more crucially, which combinations of strategies are most effective, given the positive correlation between retention rates and higher numbers of simultaneous methods employed.

“Recruitment and retention remain significant challenges for successful trial delivery. Progress has been made, but the complexity of the issues requires us to maintain momentum in evaluating current practice whilst being willing to explore new solutions,” Kearney observes. “Greater emphasis on retention is needed amongst funders, researchers and even participants to avoid recruitment targets being met at the expense of retention. 

“Fundamental to clinical research is the participant’s right to withdraw at any time,” she continues. “Research shows this right is frequently recalled and has been associated with greater willingness to participate, helping recruit people who may be initially unsure. However, withdrawal is also the most common cause of attrition across trials.”

Professor Gamble steps in: “In the same way that patients need to make an informed decision about taking part, they should also be provided with sufficient information around decisions to withdraw. What information they need to make an informed decision for withdrawal needs exploring.”

Kearney concludes: “Many patients go into research for altruistic reasons but may not be aware of the importance of retention and how withdrawal can affect a trial. The importance of retention is rarely mentioned in patient information leaflets, and better patient education may avoid recruiting patients who don’t think they will be able to remain in a trial.”

Matt Fellows