Effective recruitment of participants to a phase I study using the internet and publicity releases through charities and patient organisations: analysis of the adaptive study of IL-2 dose on regulatory T cells in type 1 diabetes (DILT1D)

Abstract

Background: A barrier to the successful development of new disease treatments is the timely recruitment of participants to experimental medicine studies that are primarily designed to investigate biological mechanisms rather than evaluate clinical efficacy. The aim of this study was to analyse the performance of three recruitment sources and the effect of publicity events during the Adaptive study of IL-2 dose on regulatory T cells in type 1 diabetes (DILT1D).

Methods: The final study outcome, demography, disease duration, residence and the effect of publicity events on the performance of three recruitment sources (clinics, type 1 diabetes (T1D) disease register and the internet) were analysed from a bespoke DILT1D recruitment database. For the internet source, the origin of website hits in relation to publicity events was also evaluated.

Results: A total of 735 potentially eligible participants were approached to identify the final 45 DILT1D participants. A total of 477 (64%) were identified via the disease register, but only 59 (12%) responded to contact. A total of 317 individuals registered with the DILT1D study team. Self-referral via the study website generated 170 (54%) registered individuals and was the most popular and successful source, with 88 (28%) sourced from diabetes clinics and 59 (19%) from the disease register. Of those with known T1D duration (N = 272), the internet and clinics sources identified a larger number (57, 21%) of newly diagnosed T1D (<100 days post-diagnosis) compared to the register (1, 0.4%). The internet extended the geographical reach of the study, enabling both national and international participation. Targeted website posts and promotional events from organisations supporting T1D research and treatment during the trial were essential to the success of the internet recruitment strategy.

Conclusions: Analysis of the DILT1D study recruitment outcomes illustrates the utility of an active internet recruitment strategy, supported by patient groups and charities, funding agencies and sponsors, in successfully conducting an early phase study in T1D. This recruitment strategy should now be evaluated in late-stage trials to develop treatments for T1D and other diseases.

Trial registration: NCT01827735 (registered: 4 April 2013); ISRCTN27852285 (registered: 23 March 2013); DRN767 (registered: 21 January 2013).

Figures

Figure 1

Gender and duration of type 1 diabetes of eligible potential participants in DILT1D. a) Males and females were equally likely to decline to participate as to enrol, with similar proportions observed across the final trial outcomes regardless of recruitment source. b) Both the clinics and internet sources were effective in identifying newly diagnosed cases of T1D (less than 100-day duration) compared to the register. A total of 19 individuals were registered at diagnosis (clinics = 12, internet = 6, register = 1). The internet source identified the largest number of cases but the clinics source was the most efficient method of locating eligible participants. Median and interquartile range shown for 246 eligible participants with less than two years duration of type 1 diabetes.

Figure 2

Consort diagram for potential DILT1D study participants registered on the DILT1D database until the point of treatment. The most successful method of registering potential participants was via self-referral from the study website (internet). Once individuals registered with the study team, similar proportions from each source proceeded to trial enrolment. Participants who were negative for autoantibodies were from the clinics (N = 2), internet (N = 2) and register (N = 1) (not permitted - outside the EU or registered after last participant treated, non-communicators - initially registered but did not respond to messages from DILT1D team).

Figure 3

Geographical spread of DILT1D study participants residence by recruitment source (N = 45). a) The internet source (blue) extended the geographical reach of the study, allowing international participation. b) The clinics (green) and register sources (red) enabled local and national recruitment. c) On average, participants recruited from the internet source (204.1 km) travelled further on average than those recruited from the clinics (54.7 km) or register (105.9 km).

Figure 4

Analysis of website traffic for the DILT1D study website. a) Website activity showing the number and origin of website hits per week and the relationship to publicity events undertaken during the trial. b) The average number of weekly website referrals before each publicity event and after, showing that the joint press release from the University of Cambridge and The Wellcome Trust had the greatest impact on website activity, with much of that increase being from science news at Wired.

Figure 5

Cumulative registration of potential DILT1D participants. Individual publicity events were observed to impact different recruitment sources with the University of Cambridge and Wellcome Trust increasing internet registration, while the Diabetes UK post increased the clinics and register sources.