SMS text pre-notification and delivery of reminder e-mails to increase response rates to postal questionnaires in the SUSPEND trial: a factorial design, randomised controlled trial

SUSRes was a randomised, controlled, 2×2 partial factorial design nested within the
SUSPEND RCT 10], testing two methods—the use of SMS text pre-notification of questionnaire delivery
and e-mail delivery of questionnaire reminders (with a link to complete the questionnaire
in a secure web site)—to improve questionnaire response rates. Within SUSPEND, patient-reported
outcomes were collected by postal questionnaire at 4 and 12 weeks, and SUSRes was
conducted in 24 SUSPEND sites throughout the UK from June 2012 to March 2014. SUSRes
was not planned at the outset of the SUSPEND RCT; rather, it was developed in response
to lower-than-expected response rates to postal questionnaires within SUSPEND.

Ethics and other approvals

The study was reviewed and approved by the East of Scotland Research Ethics Service
as a substantial amendment to the SUSPEND trial (Research Ethics Committee) reference
10/S0501/31) and received approval from local research and development departments
at the SUSPEND sites (Aberdeen Royal Infirmary; Addenbrooke’s Hospital, Cambridge;
Bristol Royal Infirmary; Broadgreen Hospital, Liverpool; Cheltenham General Hospital;
Derriford Hospital, Plymouth; Freeman Hospital, Newcastle upon Tyne; Guy’s Hospital,
London; Manchester Royal Infirmary; Morriston Hospital, Swansea; Norfolk and Norwich
University Hospital; Pinderfields Hospital, Wakefield; Queen Elizabeth Hospital, Birmingham;
Raigmore Hospital, Inverness; Royal Hallamshire Hospital, Sheffield; Southampton General
Hospital; Southmead Hospital, Bristol; St George’s Hospital, London; St James’s University
Hospital, Leeds; Sunderland Royal Hospital; The James Cook University Hospital, Middlesbrough;
Torbay Hospital, Torquay; University Hospital of South Manchester; Western General
Hospital, Edinburgh) before commencing. Participants provided written informed consent
for SUSPEND which encompassed the SUSRes study.

Participants

Participants who were newly randomly assigned to the SUSPEND trial, had not reached
the 4-week time point, and were willing to supply a mobile phone number or an e-mail
address (or both) were considered for the study. Participants could be included in
one or both comparisons. All participants who met the inclusion criteria and where
a mobile phone number was recorded were included in the SMS text comparison. Participants
for whom an e-mail address was provided were included in the e-mail comparison. All
participants provided written informed consent for the SUSPEND trial and therefore
this study.

Interventions and controls

There were two comparisons (each with a separate randomisation) within the SUSRes
study.

SMS text pre-notification comparison

All participants randomly assigned to the intervention arm were sent an SMS text message
pre-notification of the delivery of the initial 4- and 12-week questionnaires. The
message was generated automatically from the SUSPEND trial database and sent via an
external supplier the same day the questionnaire was dispatched from the trial office.
The SMS read: “Many thanks for participating in the SUSPEND trial. You will shortly
receive your xx-week questionnaire. We hope you can take a few minutes to complete
this and return it to us”. Participants randomly assigned to the control arm were
not sent any pre-notification of the delivery of the 4- and 12-week questionnaires.

E-mail reminder comparison

All participants who were randomly assigned to the intervention arm and who did not
respond to the initial 4- or 12-week questionnaire received an e-mail which included
a link to complete the questionnaire online or were invited to return the paper copy
if they wished. Participants who were randomly assigned to the control arm and who
did not respond to the initial 4- or 12-week questionnaire received their reminder
by post with a further copy of the questionnaire.

Both types of reminder would have been generated on the same day, two weeks after
the initial questionnaire was sent out by post from the trial office, by the SUSPEND
trial database. The e-mails were automatically sent from the trial database and the
postal reminders were printed and posted by trial office staff on the same day.

The wording of the reminder e-mail reflected the wording and layout of the letter
that accompanied the postal questionnaires to avoid any bias by modifying this. The
online questionnaire contained the same questions as the postal questionnaire.

Allocation and randomisation

Participants were randomly allocated to the intervention or control groups of the
SUSRes study on a 1:1 basis by using a computer-generated system that was concealed
and remote from the users. The randomisation algorithm was permuted blocks (block
size of 4) stratified by age (? 40 years or 40 years) and sex as these variables
are known to affect response rate 8], 9]. The allocated group applied to both SUSPEND questionnaire time points (4 and 12
weeks). Owing to the nature of the intervention, it was not possible to blind the
participants or trial office staff to allocation; however, the researchers remained
blind.

Outcomes

The primary outcome for both comparisons was defined as questionnaire response rate
at each time point. There were no secondary outcomes.

Sample size

The sample size for the SUSRes study was dictated by the number of participants still
to be recruited into SUSPEND at the time the study started (710); therefore, no formal
sample size calculation was undertaken.

Statistical analysis

All analyses were performed in Microsoft Excel (2010; Microsoft Corporation, Redmond,
WA, USA) and IBM SPSS statistics version 22.0 (IBM Corporation, Somers, NY, USA).
Two comparisons were made: SMS text message pre-notification versus no pre-notification
of questionnaire delivery and questionnaire e-mail reminder versus postal reminder.

Participant baseline characteristics (age and sex) between the intervention and control
groups within each intervention comparison were compared by using an independent t test or chi-squared test as appropriate.

The primary outcomes (response rate at each time point) were analysed on the basis
of the intention-to-treat principle 11], and all participants were analysed as randomly assigned. To assess the impact of
each intervention, adjusted odds ratios (ORs) (adjusted for the stratification variables
age and sex) with 95 % confidence intervals (CIs) were computed by using logistic
regression. Allocation to the other comparison was also added as a co-variate in the
adjusted analysis. Possible interaction between the two interventions was explored
with logistic regression analysis in those participants who had been randomly assigned
in both comparisons.

The effects of age and sex were considered in pre-specified subgroup analyses by using
the same analysis techniques described above and adjusted for age or sex as appropriate
and allocation to the other comparison.

The CONSORT (Consolidated Standards Of Reporting Trials) checklist 12] for this study can be found in Additional file 1.