Opt-out rates and reasons for non-participation in a single-arm feasibility trial (ENGAGE) of a guided internet-administered CBT-based intervention for parents of children treated for cancer: a nested cross-sectional survey

Objectives Difficulties with recruitment into clinical trials are common. An opt-out recruitment strategy, whereby potential participants can decline further contact about a study (opt-out), and non-responders are contacted, may facilitate participation. Primary objectives examined opt-out and consent rates, mode and time point of opt-out, and sociodemographic characteristics of those who opted out versus those who chose to participate in a single-arm feasibility trial (ENGAGE) of a guided, internet-administered, cognitive–behavioural therapy-based intervention for parents of children treated for cancer. Secondary objectives examined reasons for non-participation. Design A cross-sectional survey nested within the ENGAGE feasibility trial. Setting The intervention was delivered from Uppsala University, with parents located throughout Sweden. Participants Potential participants were recruited 3 months–5 years following their child ending treatment for cancer and were identified via their personal identification number (via the Swedish Childhood Cancer Registry and Swedish Tax Agency) and invited via postal invitation packs and could opt out via post, online, telephone or email. Those who did not opt out or consent, within 4 weeks, received up to five telephone calls and/or one postal reminder. Results Of 509 invited, 164 (32.2%) opted out, 78 (47.6%) via post, 53 (32.3%) via telephone, 24 (14.6%) online, and 6 (3.7%) via email, 88 (53.7%) opted out after at least one telephone call and/or postal reminder. There was a trend for parents with lower educational levels to opt out. No need of psychological support, lack of time, and no interest in internet-administered self-help were frequently reported reasons for non-participation. Conclusions Results emphasise the importance of using different opt-out modes and suggest future research should consider how to widen study participation for parents with lower education levels. Self-identifying a need for psychological support and the acceptability of internet-administered self-help are important factors for participation and should be considered in future research to increase recruitment. Trial registration number ISRCTN57233429.


Introduction
Good background and rational to the study provided.

Methods
Repeatable methods describe, would be interesting to know whether the use of personalised invitation letters effects the number of non-responders. e.g. fewer opt-out or more respond to opt-out without the need subsequent reminders before opting out.

Results and discussion
Results clearly presented and reflected in the discussion. Identifies limitations for the study and suggests future research direction.

REVIEWER
Liu, Na The Second Affiliated Hospital of Chongqing Medical University REVIEW RETURNED 07-Feb-2022

GENERAL COMMENTS
Thank you very much for inviting me to make suggestions on this manuscript. The manuscript research design is rigorous, logical, accurate and comprehensive in analyzing the reasons for participants' participation and withdrawal. However, the primary objective of this paper is opt-out modes, but the results and discussion of the paper lack further explanation of the differences between different opt-out modes and the potential significance of choosing different opt-out modes. What practical guiding significance does this study have for future research? Please explain further. Thank-you for this recommendation. We have added a very brief background to the Abstract (see page 2):

VERSION 1 -AUTHOR RESPONSE
Difficulties with recruitment into clinical trials are common. An opt-out recruitment strategy, whereby potential participants can decline further contact about a study (opt-out), and nonresponders are contacted, may facilitate participation.

Introduction Good background and rational to the study provided.
Thank-you for this feedback.

Methods
Repeatable methods describe, would be interesting to know whether the use of personalized invitation letters effects the number of non-responders. e.g. fewer opt-out or more respond to opt-out without the need subsequent reminders before opting out. Thank-you for this comment. We have submitted the study-within-a-trial (SWAT) as a separate manuscript, have recently received reviewers' comments and the manuscript has been revised and resubmitted, and we hope will be accepted for publication soon. One of the secondary outcomes was the proportion of potential participants invited into the trial in each group that opted out of the ENGAGE host feasibility trial. The number for opting out of the trial was smaller in the personalized invitation letter group than the non-personalized invitation letter group. However, we would prefer to only report this data in the SWAT manuscript and have informed the journal that this data is not presented for publication elsewhere. As such, it is not possible to add this data to the manuscript.

Results and discussion
Results clearly presented and reflected in the discussion. Identifies limitations for the study and suggests future research direction. Thank-you for this comment.

Reviewer: 2 Miss Na Liu, The Second Affiliated Hospital of Chongqing Medical University
Comments to the Author: Thank you very much for inviting me to make suggestions on this manuscript. The manuscript research design is rigorous, logical, accurate and comprehensive in analyzing the reasons for participants' participation and withdrawal. However, the primary objective of this paper is optout modes, but the results and discussion of the paper lack further explanation of the differences between different opt-out modes and the potential significance of choosing different opt-out modes. What practical guiding significance does this study have for future research? Please explain further.
Thank-you for your positive comments on the manuscript. We have four primary objectives, to examine opt-out and consent rates, mode and time point of opt-out, and sociodemographic characteristics of those who opted out versus chose to participate in the ENGAGE feasibility trial. At present, we have discussed findings relating to each of these objectives. We have made some minor amendments to the wording of the study objectives, to make the four primary objectives clear (see To the best of our knowledge, this is the first study to specifically examine the rate of opt-out across four different modes of opt-out. Commonly, studies utilising opt-out recruitment strategies include only one mode of opt-out, for example via the post [26; 58] or telephone. [59] The most frequent mode of opt-out in the ENGAGE feasibility trial was paper via the post (47.6 %), followed by telephone (32.3%), with findings similar to a quantitative telephone survey with military veterans examining rural and urban differences in attitudes towards mental health care and influence on mental health service use, whereby opt-out via the post was more frequent (75%) than opt-out via the telephone (25% out via the post as implying higher levels of security and privacy than using an electronic personal identification method (BankID) to opt-out online. However, the utilisation of different modes of opt-out was easy to implement and given the variation in response rate across the four opt-out modes, future studies may wish to offer potential participants multiple modes of opt-out. Further, future studies could examine additional potential modes of opt-out, for example, via short message service (SMS). [65] We have also added further amendments to the Conclusion section regarding future research (see page 19): This is the first study to examine the use of an opt-out strategy and different modes of opt-out within a feasibility trial examining the feasibility and acceptability of an internet-administered, guided, LICBT based self-help intervention for parents of children treated for cancer. Though recruitment into clinical trials is challenging, projected recruitment rates were exceeded.
Despite the most frequent mode of opt-out being paper via the post, given the implementation of different modes of opt-out was straightforward, coupled with the variation in response rate across the four opt-out modes, future studies should consider using multiple modes of opt-out.