Introduction
In March 2013, NHS England extended its national Bowel Cancer Screening Programme
to include once-only flexible sigmoidoscopy screening for men and women aged 55 years
[1]. The decision to extend the program and include flexible sigmoidoscopy was made
in response to the results of a large UK randomized controlled trial (RCT), which
demonstrated that a one-off screen between the ages of 55 and 64 years significantly
reduced the incidence and mortality of colorectal cancer (CRC) among screened adults
[2].
Uptake of the test (also referred to as bowel scope screening [BSS]) in England is
currently very low (only 43 % of invitees attend an appointment) [3], which will ultimately undermine the clinical effectiveness of the program [4]. Data on uptake in other countries are sparse, but it has been documented to be
as low as 29 % in regions of Italy [5] and as high as 63 % in the Norwegian Colorectal Cancer Prevention Trial [6].
Detailed surveys issued to nonparticipants and their healthcare providers have identified
the most prevalent patient-related factors for not attending a flexible sigmoidoscopy
appointment as: a lack of current health problems, practical barriers (i. e. inconvenient
appointment time/day), worry about pain, discomfort, or injury associated with the
examination, and not wanting to know about any health issues [7]
[8]
[9]. Subsequent studies have suggested that interventions which address patient-specific
barriers to the test might improve uptake [10]
[11]; however, the evidence to support the use of such strategies is inconsistent [12]
[13]
[14]
[15]
[16]
[17]
[18].
Organizational changes, such as those affecting the mechanisms for scheduling appointments,
self-referral, and patient outreach have been more effective at increasing patient
participation [19]
[20]
[21]
[22]
[23]. Pre-notification letters, timed appointments, and pre-appointment reminders have
all been shown to improve uptake of flexible sigmoidoscopy screening and, as with
other screening, these specific strategies have been incorporated into the BSS pathway
[5]
[24]
[25]
[26]. Repeat invitations have additionally been shown to improve uptake of CRC screening
tests (such as the fecal occult blood test and colonoscopy) by providing additional
opportunities for people to take part [27]
[28]
[29]. In Italy, Senore et al. made similar observations in the context of a flexible
sigmoidoscopy-based program when they sent former nonresponders a reminder letter
3 months after the invitation, yielding a 4.5 % absolute increase in uptake [5].
In the BSS program, invitees who do not attend screening at the age of 55 years can
self-refer for the test up until the age of 60, but do not receive a formal reminder
of this opportunity. A recent single-arm feasibility study examining the format of
a 12-month reminder for a center in London found that, when sent with a bespoke cover
letter and information leaflet, the reminder facilitated uptake in 15.5 % of former
nonresponders [30]. Although highly promising, these current studies of nonresponder reminders are
limited to observational data from single-arm trials [5]
[30]. The effectiveness of these interventions and some of their key components therefore
still require formal evaluation in multi-arm RCTs. One such component requiring further
investigation is the bespoke, theory-based leaflet used in the feasibility study [30], which had been designed to address several barriers to screening by providing bespoke
information on how to get to the screening center and by including testimonials from
locally screened adults [30]. Using local tailoring for the whole program would incur additional financial and
logistical costs, which would need to be justified.
Another hitherto untested component of these reminders is their differential impact
on specific groups of nonparticipants, as to date current studies have focused exclusively
on former nonresponders (i. e. individuals who previously did not confirm nor attend
their appointment) [5]
[30]. Previous research has shown that individuals who confirm an appointment, but then
do not attend (i. e. former nonattenders), are both demographically and qualitatively
different from their screened and nonresponding counterparts [31], with previously screened adults having “better subjective health” and “lower levels
of deprivation” and former nonresponders having “lower levels of perceived benefits
and anticipated regret” and higher levels of “cancer fear and fatalism” [31]. As it stands, there is no published information about the prevalence of this important
subgroup of nonparticipants in the English BSS program or how they would respond to
interventions inviting them to re-engage with the program.
The present study therefore set out to extend the evaluation of nonparticipant reminders
to increase uptake of flexible sigmoidoscopy screening by: 1) comparing uptake between
individuals receiving a reminder compared with usual care (i. e. no reminder) in an
RCT; 2) subdividing nonparticipants into former nonattenders and former nonresponders;
and 3) testing the added benefit of including a bespoke theory-based leaflet with
the reminder.
Patients and methods
Study design and trial setting
We performed a single-blind RCT with three parallel arms in the London Boroughs of
Brent and Harrow. Researchers were blinded to the treatment that subjects received
until all data had been collected at the end of the study. Because individuals were
given a reminder plus the standard information booklet, no reminder, or a reminder
plus the locally tailored theory-based leaflet, it was not possible to blind them
to the treatment they received. In terms of the study setting, the London Boroughs
of Brent and Harrow have below-average uptake (40 % vs. 43 %, respectively), and individuals
living in these boroughs predominantly live in the most ethnically diverse and socioeconomically
deprived areas of England [4].
Study population
Eligible adults were men and women registered with a general practice in the London
Boroughs of Brent and Harrow. Adults registered with these practices were eligible
for inclusion in the study if they had not attended a BSS appointment within 12 months
of receiving their invitation. Eligible adults included both those who previously
did not respond to the initial invitation at the age of 55 years (former nonresponders)
as well as those who confirmed an appointment but did not attend (former nonattenders).
Procedures
Eligible adults were identified from the NHS Bowel Cancer Screening System (BCSS)
[32]. To ensure workforce capacity (i. e. that all self-referred appointments could be
facilitated without disruption to routine appointments), eligible adults were enrolled
over a 20-week period spanning February to August 2015. On the basis that there was
capacity to facilitate an additional five appointments per week, and that 5 % of individuals
across the three groups would make and attend an appointment each week, we conservatively
selected 69 adults for inclusion in the study each week (from a variable weekly total)
using simple pseudo-random selection methods [33] ([Fig. 1]). Individuals selected for inclusion in the study were then randomly assigned using
simple pseudorandom allocation methods (in a 1:1:1 ratio) to receive: no reminder
(control); a mailed 12-month reminder plus a standard information booklet (TMR-SIB);
or a mailed 12-month reminder plus a locally tailored theory-based leaflet (TMR-TBL)
designed to address barriers to screening.
Fig. 1 Trial flowchart/CONSORT diagram. BCSS, Bowel Cancer Screening System; TMR, 12-month
reminder; SIB, standard information booklet; TBL, theory-based leaflet.
Individuals allocated to the control group were provided with usual care and therefore
did not receive a reminder. Individuals allocated to the reminder groups were sent
a reminder letter with one of two leaflets, an appointment-request slip, and a freepost
return envelope addressed to St Mark’s Bowel Cancer Screening Centre ([Fig. 1]).
Individuals in both reminder groups were able to book an appointment by returning
their appointment-request slip in the freepost envelope provided, thereby initiating
a call from a member of the administrative team to arrange an appointment, or by calling
the screening center directly on the Freephone telephone number highlighted in the
reminder letter.
Individuals not responding to the reminder within 4 weeks were sent a follow-up reminder
([Fig. 1]), which also included an appointment-request slip, the allocated information leaflet,
and a freepost return envelope. Individuals were then given 8 more weeks to respond,
after which they were not included in the study results. Individuals who self-referred
for the test also received a pre-appointment text-message reminder and telephone call,
as per routine appointments at St Mark’s Hospital.
The study was approved by the North East – Tyne & Wear South Research Ethics Service
(Ref: 15 /NE/0043) and was registered with the International Standard Randomised Controlled
Trials Number Registry for transparency (trial ID: ISRCTN44293755).
Intervention details
Full descriptions of the intervention materials and their development are available
in the feasibility study [30].
12-month reminder
The 12-month reminder was a personally addressed letter from St Mark’s Hospital that
invited recipients to make a new appointment by returning an appointment-request slip
or by calling the Freephone telephone number for the screening center ([Appendix e1], available online). The reminder also gave recipients the option to express a preference
for the day and time of the appointment, as well as the sex of the practitioner performing
the test.
Appendix e1 Reminder letter sent to randomized former nonresponders and nonattenders.
Theory-based leaflet (TMR-TBL)
The theory-based leaflet was a locally tailored leaflet co-designed by Resonant, a
social marketing company specializing in health behavior (see [Appendix e2], available online). The development of the leaflet was based on two psychological
models of behavior previously used to explain factors associated with uptake [34]: the Health Belief Model [35] and Social Cognitive Theory [36]. To address barriers to participation, the leaflet included an educational/knowledge-building
component and several practical components designed to improve self-efficacy (including
local transport information and directions to the hospital).
Appendix e2 Theory-based leaflet sent to randomized former nonresponders and nonattenders.
Standard information booklet
The standard information booklet was the same 16-page booklet that was previously
sent with the initial invitation as part of the national program (available from:
https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/423928 /bowel-scope-screening.pdf). The standard information booklet was developed by King’s Health Partners who developed
the booklet in accordance with principles put forth by NHS England’s informed choice
initiative [37].
Follow-up reminder
The follow-up reminder was a personally addressed letter from St Mark’s Hospital that
reiterated the opportunity to self-refer for screening up until the age of 60 years
([Appendix e3], available online).
Appendix e3 Follow-up reminder letter.
Measures
Routinely available data on the BCSS were used to verify self-referral and attendance
4 and 12 weeks following the distribution of the 12-month reminder letter. The BCSS
was also consulted to obtain the sex, area (i. e. Brent or Harrow), and initial episode
status (i. e. former nonresponder or former nonattender) of each person included in
the trial. For those who attended an appointment, the BCSS was additionally consulted
to obtain the proportion of people screened who had one or more adenomas detected.
An area-based socioeconomic deprivation score was generated for each person by converting
individual postcodes into a score on the 2010 Index of Multiple Deprivation (IMD)
[38]. Area-level IMD scores were then categorized into tertiles of their regional distributions
to enable comparisons between the most and least deprived areas.
Statistics
Sample size
The sample size (n = 1383) was calculated using a standard test of difference between
two proportions. As the study included three trial arms (one receiving usual care
and two receiving a reminder with one of two leaflets), the calculation was repeated
for each pairwise comparison comprising a primary research question in the planned
analysis. The final calculation gave a total sample size requirement of 461 people
per trial arm to test for a 5 % difference in uptake between any two of the three
groups, with expected values of 0 %, 5 %, and 10 % for the control, TMR-SIB, and TMR-TBL
groups respectively. The study was designed to detect differences at the two-sided
5 % alpha level with a 20 % margin for type II error.
Data analysis
Univariate logistic regression was used to investigate the associations between treatment
groups and self-referral and uptake. To counteract the problem of multiple comparisons,
we used the Bonferroni correction method, comparing outcomes to an adjusted significance
level of 0.015. Adjusted odds ratios (ORs) and 95 % confidence intervals (CIs) were
calculated using multivariate logistic regression after adjusting for baseline characteristics.
To explore possible associations between not attending a confirmed appointment and
previous episode status (i. e. former nonattender, former nonresponder), we carried
out a subgroup analysis using univariate and multivariate logistic regression.
The adenoma detection rate (ADR) was reported using descriptive statistics; the data
were analyzed on an intention-to-treat basis using SPSS software (version 22; IBM
Corp., Armonk, New York, USA).
Results
Sample characteristics
The study took place between February and August 2015, with follow-up until October
2015. A total of 1383 adults were randomized and analyzed as allocated. The majority
of individuals were registered with a general practice in the London Borough of Brent
(n = 928; 67.1 %), did not respond to the initial invitation (n = 1255, 90.7 %), and
were female (n = 727; 52.6 %) ([Table 1]).
Table 1
Description of the trial population.
|
Control
(n = 461)
|
TMR-SIB
(n = 461)
|
TMR-TBL
(n = 461)
|
Total
(n = 1383)
|
|
Sex, n (%)
|
|
Female
|
261 (56.6)
|
238 (51.6)
|
228 (49.5)
|
727 (52.6)
|
|
Male
|
200 (43.4)
|
223 (48.4)
|
233 (50.5)
|
656 (47.4)
|
|
Area, n (%)
|
|
Brent
|
304 (65.9)
|
302 (65.5)
|
322 (69.8)
|
928 (67.1)
|
|
Harrow
|
157 (34.1)
|
159 (34.5)
|
139 (30.2)
|
455 (32.9)
|
|
Tertile of deprivation (IMD score), n (%)
|
|
Tertile 1 (0.00 – 17.68)
|
152 (33.0)
|
144 (31.2)
|
133 (28.9)
|
429 (31.0)
|
|
Tertile 2 (17.69 – 27.50)
|
164 (35.5)
|
162 (35.0)
|
179 (38.8)
|
505 (36.5)
|
|
Tertile 3 (27.51 – 80)
|
140 (30.4)
|
151 (32.8)
|
144 (31.2)
|
435 (31.5)
|
|
Missing
|
5 (1.1)
|
4 (0.9)
|
5 (1.1)
|
14 (1.0)
|
|
Initial episode status, n (%)
|
|
Nonresponder
|
411 (89.2)
|
408 (88.5)
|
436 (94.6)
|
1255 (90.7)
|
|
Nonattender
|
50 (10.8)
|
53 (11.5)
|
25 (6.4)
|
128 (9.3)
|
TMR, 12-month reminder; SIB, standard information booklet; TBL, theory-based leaflet;
IMD, Index of Multiple Deprivation.
Uptake of bowel scope screening
In total, 119 people (8.6 %) attended an appointment across all three study groups
([Table 2]). A further 41 (3.0 %) made an appointment, but then either did not attend (n = 21)
or cancelled (n = 20), leaving 1223 (88.4 %) adult men and women who neither made
nor attended an appointment ([Table 2]).
Table 2
Self-referral and uptake by trial arm (univariate and multivariate regression outcomes).
|
Mean, n (%)
|
Unadjusted OR (95 %CI)
|
Adjusted OR (95 %CI)[1]
|
|
Made an appointment (n = 160)
|
|
Control vs. TMR-SIB
|
1 vs. 64 (0.2 vs. 13.9)
|
74.16 (10.24 – 536.97) [2]
|
73.27 (10.11 – 531.11) [2]
|
|
Control vs. TMR-TBL
|
1 vs. 95 (0.2 vs. 20.6)
|
119.40 (16.57 – 860.49) [2]
|
130.36 (18.05 – 941.54) [2]
|
|
TMR-SIB vs. TMR-TBL
|
64 vs. 95 (13.9 vs. 20.6)
|
1.61 (1.14 – 2.28) [3]
|
1.78 (1.25 – 2.54) [3]
|
|
Attended an appointment (n = 119)
|
|
Control vs. TMR-SIB
|
1 vs. 48 (0.2 vs. 10.4)
|
53.46 (7.35 – 389.05) [2]
|
53.73 (7.38 – 391.39) [2]
|
|
Control vs. TMR-TBL
|
1 vs. 70 (0.2 vs. 15.2)
|
82.35 (11.39 – 595.58) [2]
|
89.01 (12.28 – 645.40) [2]
|
|
TMR-SIB vs. TMR-TBL
|
48 vs. 70 (10.4 vs. 15.2)
|
1.54 (1.04 – 2.28)[4]
|
1.69 (1.13 – 2.52) [3]
|
OR, odds ratio; CI, confidence interval. n = 461 for all groups, respectively.
1 Adjusted ORs and 95 %CIs are adjusted for sex, area, deprivation, and initial episode
status.
2
P ≤ 0.001
3
P ≤ 0.01
4
P ≤ 0.05
There was strong evidence of differences in booked and attended appointments between
the reminder groups and the control ([Table 2]). A total of 48 individuals (10.4 %) in the TMR-SIB group and 70 (15.2 %) in the
TMR-TBL group attended an appointment ([Table 2]) compared with only 1 (0.2 %) in the control group (OR 53.5, 95 %CIs 7.4 – 389.1,
P < 0.001; OR 82.4, 95 %CIs 11.4 – 595.6, P < 0.001 for the TMR-SIB and TMR-TBL groups, respectively). There was also a strong
trend toward differences in uptake between the reminder groups, with individuals in
the TMR-TBL group being more likely to attend an appointment than individuals in the
TMR-SIB group (OR 1.5, 95 %CI 1.0 – 2.3, P = 0.03).
Results were similar after adjusting for baseline characteristics in the multivariate
analysis ([Table 2]), with strong evidence of significant differences between the reminder groups and
control (TMR-SIB vs. control: OR 53.7, 95 %CIs 7.4 – 391.4, P < 0.001; TMR-TBL vs. control: OR 89.0, 95 %CIs 12.3 – 645.4, P < 0.001). After adjusting for baseline characteristics, there was also strong evidence
for a difference in participation between intervention groups, with individuals in
the TMR-TBL group being more likely to book and attend an appointment than individuals
in the TMR-SIB group (OR 1.7, 95 %CIs 1.1 – 2.5; P = 0.01). There was also strong evidence of a difference in uptake by initial episode
status after adjusting for study group and other baseline characteristics ([Table 3]), with former nonattenders being nearly twice as likely to book and attend an appointment
than former nonresponders (14.2 % and 8.0 %, respectively; OR 2.5, 95 %CI 1.4 – 4.4;
P < 0.01). There was no evidence of an association between screening uptake and sex,
regional IMD tertile, or area (all P values > 0.05; [Table 3]).
Table 3
Self-referral and uptake by baseline characteristics (univariate and multivariate
regression outcomes).
|
Made an appointment,
n (%)
|
Unadjusted OR
(95 %CI)
|
Adjusted OR
(95 %CI)[1]
|
Attended an appointment,
n (%)
|
Unadjusted OR
(95 %CI)
|
Adjusted OR
(95 %CI)[1]
|
|
Sex
|
|
Women[2] (n = 727)
|
83 (11.4)
|
–
|
–
|
57 (7.8)
|
–
|
–
|
|
Men (n = 656)
|
77 (11.7)
|
1.03 (0.74 – 1.44)
|
0.94 (0.67 – 1.33)
|
62 (9.5)
|
1.23 (0.84 – 1.79)
|
1.18 (0.80 – 1.75)
|
|
Area
|
|
Brent[2] (n = 926)
|
101 (10.9)
|
–
|
–
|
75 (8.1)
|
–
|
–
|
|
Harrow (n = 457)
|
59 (12.9)
|
1.21 (0.86 – 1.71)
|
1.39 (0.89 – 2.18)
|
44 (9.6)
|
1.21 (0.82 – 1.79)
|
1.34 (0.80 – 2.22)
|
|
Deprivation
|
|
Tertile 1[2] (n = 429)
|
49 (11.4)
|
–
|
–
|
38 (8.9)
|
–
|
–
|
|
Tertile 2 (n = 505)
|
59 (11.7)
|
1.03 (0.69 – 1.54)
|
1.20 (0.74 – 1.95)
|
41 (8.1)
|
0.91 (0.57 – 1.44)
|
1.05 (0.61 – 1.81)
|
|
Tertile 3 (n = 435)
|
51 (11.7)
|
1.03 (0.68 – 1.56)
|
1.29 (0.76 – 2.21)
|
40 (9.2)
|
1.04 (0.65 – 1.66)
|
1.28 (0.71 – 2.33)
|
|
Initial episode status
|
|
Nonresponder[2] (n = 1256)
|
138 (11.0)
|
–
|
–
|
101 (8.0)
|
–
|
–
|
|
Nonattender (n = 127)
|
22 (17.3)
|
1.70[4] (1.04 – 2.78)
|
2.24[3] (1.30 – 3.85)
|
18 (14.2)
|
1.89[4] (1.10 – 3.24)
|
2.45[3] (1.36 – 4.40)
|
OR, odds ratio; CI, confidence interval.
1 Adjusted ORs and 95 %CIs are adjusted for trial arm and all other covariates in the
table.
2 Reference category.
3
P ≤ 0.01
4
P ≤ 0.05
Confirmed appointments
A total of 41 individuals booked an appointment but did not attend. Attendance of
a confirmed appointment was higher among former nonattenders than former nonresponders
(81.8 % vs. 73.0 %); however, the results of the regression revealed that there was
no significant difference between these two groups (OR 1.4, 95 %CIs 0.4 – 4.6) ([Appendix e4]). A significant difference in attendance was observed between men and women (80.5 %
vs. 68.3 %), with men being more likely to attend than their female counterparts (OR
2.2, 95 %CI 1.0 – 4.7). There were no significant differences in the nonattendance
rate for any of the other covariates included in the analysis (all P values > 0.05).
Appendix e4
Uptake of self-referred appointment by baseline characteristics and trial arm (univariate
and multivariate regression).
|
Attended appointment/self-referred, n (%)
|
Unadjusted OR (95 %CI)
|
Adjusted OR (95 %CI)
|
|
Treatment group comparisons
|
|
TMR-SIB[1] (n = 64)
|
48 (75.0)
|
–
|
–
|
|
TMR-TBL[1] (n = 95)
|
70 (73.7)
|
0.93 (0.45 – 1.93)
|
0.93 (0.43 – 1.98)
|
|
Sex
|
|
Women[1] (n = 82)
|
56 (68.3)
|
–
|
–
|
|
Men (n = 77)
|
62 (80.5)
|
1.92 (0.92 – 3.99)
|
2.17 (1.00 – 4.67)
|
|
Area
|
|
Brent[1] (n = 100)
|
74 (74.0)
|
–
|
–
|
|
Harrow (n = 59)
|
44 (74.6)
|
1.03 (0.49 – 2.15)
|
0.89 (0.35 – 2.25)
|
|
Deprivation
|
|
Tertile 1[1] (n = 49)
|
38 (77.6)
|
–
|
–
|
|
Tertile 2 (n = 59)
|
41 (69.5)
|
0.66 (0.28 – 1.57)
|
0.58 (0.21 – 1.59)
|
|
Tertile 3 (n = 50)
|
39 (78.0)
|
1.03 (0.40 – 2.65)
|
1.02 (0.32 – 3.31)
|
|
Initial episode status
|
|
Nonresponder[1] (n = 137)
|
100 (73.0)
|
–
|
–
|
|
Nonattender (n = 22)
|
18 (81.8)
|
1.67 (0.53 – 5.24)
|
1.42 (0.43 – 4.61)
|
OR, odds ratio; CI, confidence interval.
Table excludes data from the control arm which included one self-referred appointment
only.
1 Reference category
Adenoma detection rate
Of the 119 individuals who attended an appointment and were screened, 9 (7.6 %) had
one or more adenomas detected, 6 of whom also met the clinical criteria for colonoscopy
and subsequently underwent further examination. No patient was diagnosed with cancer.
Discussion
This trial was initiated to test the impact of a 12-month reminder compared with usual
care in an RCT. It is the first study to subdivide nonparticipants in order to test
the effect of long-term reminders on former nonresponders and former nonattenders
independently. This study also examined the added benefit of including a bespoke,
theory-based leaflet with the reminder by including a third parallel arm to the trial
design.
The results of this RCT provide strong evidence to support the use of a 12-month reminder
in the national BSS program and highlight an additional benefit of including a bespoke,
theory-based leaflet among a group of adults who have previously received the full
suite of information as part of the initial invitation (uptake was 0.2 %, 10.4 % and
15.2 % in the control, TMR-SIB, and TMR-TBL groups, respectively).
At the current rate of attendance (43 %) [3], inclusion of a 12-month reminder in the national BSS program would increase uptake
by approximately 6 – 9 percentage points (estimated by multiplying the proportion
of adults not attending an initial appointment [0.57] by the proportion of adults
attending in response to the 12-month reminder either with the SIB [0.10] or the TBL
[0.15]), depending on which of the two leaflets were adopted. As uptake was consistent
between men and women, as well as between tertiles of socioeconomic deprivation, it
is unlikely that implementing a 12-month reminder with either leaflet would exacerbate
existing inequalities in participation [3].
It is interesting to note that while uptake did not vary by sex, area, or area-level
deprivation, it did vary by previous episode status, with former nonattenders being
more likely to book and attend an appointment than former nonresponders (14.2 % vs.
8.0 %). One possible explanation for this difference is that former nonattenders,
who perceive fewer barriers and more benefits to screening than nonresponders, are
qualitatively similar to screened adults, but have difficulty translating their intentions
into actions due to circumstantial aspects, such as poor health [31]. Previous research by Ferrer et al. has shown that participation in CRC screening
is a behavioral process comprising several qualitatively distinct stages through which
individuals move based on their readiness to be screened [39]. Each stage is strongly associated with a specific set of attitudes and beliefs
toward the test, and it may be that the interventions used in our study were more
effective at facilitating forward-stage transitions in former nonattenders by addressing
issues that are specific to those who have already engaged with the program by making
an appointment. It is also possible that it may simply be easier to facilitate forward-stage
transitions in individuals who have previously responded to the initial invitation
than in individuals who have not, and so the higher response in former nonattenders
may have been observed for this reason. It is also important to note that, contrary
to previous studies suggesting that nonattenders might never translate their intentions
into actions [30], we found that the attendance rate for those confirming an appointment was actually
higher (although not significantly higher) among former nonattenders than former nonresponders
(81.2 % and 73.0 %, respectively).
Although this particular subgroup of nonparticipants does not comprise a substantial
proportion of the population (we find that they only account for approximately 10 %
of individuals), they do represent a willing group who respond well to these interventions
and as such should not be excluded.
We also found that women who made an appointment were less likely to attend screening
than men who made an appointment (68.3 % vs. 80.5 %), and this was consistent with
previous research examining factors associated with non-attendance [31]. One possible explanation as to why women were less likely to attend a screening
appointment than men who made an appointment is that women perceive more barriers
to (flexible sigmoidoscopy) screening [40], which make it more difficult for them to attend [8].
It is possible that a telephonic reminder would have been more effective [19]
[20]
[21]. However, telephonic reminders are not considered cost-effective for CRC screening
and as such are not recommended by the European Quality Assurance Guidelines for Colorectal
Cancer Screening [41]. In addition, the screening center does not have access to patient telephone numbers
(unless patients have provided a number, in which case they are available for responders
and a telephone reminder for the appointment is given). Given this and the results
of the present study, we would advocate the adoption of a 12-month mailed reminder
for nonparticipants by the national program prior to full population coverage in 2018. From
a clinical perspective, this intervention can be considered worthwhile, given that
the ADR reported in our study (7.6 %) is comparable to the rate associated with the
initial invitation (9.8 %) [24].
There has been little positive research concerning the impact of theory-based materials
on CRC screening rates [12]
[13]
[14]
[15]
[16]
[17]
[18], particularly with regard to flexible sigmoidoscopy screening [21]. The finding that the bespoke, theory-based leaflet used in this study was effective
is therefore highly promising. Not only does it demonstrate that such materials can
be effective, but also that they can be implemented in ways that do not contravene
General Medical Council guidelines for informed consent. It is possible that this
type of leaflet might also be effective at other stages of the invitation pathway
(for instance, if sent with the pre-notification or pre-appointment reminder letters),
although it may be that it worked well because it was sent to a group of individuals
for whom the standard information was not suitable.
Our study has several limitations. First, we only tested the impact of a 12-month
reminder at one screening center and cannot say whether the interventions described
here would be as effective outside of the study setting. Furthermore, to ensure endoscopy
capacity (i. e. that all appointments could be facilitated), we selected only a proportion
of former nonparticipants for this trial and not the entire eligible population. At
this point it would be important to investigate the feasibility of rolling out these
reminders across the entire eligible population, as well as at other centers. Other
factors, such as recent bowel symptoms [8], which are associated with uptake of BSS, could not be examined because our study
was embedded within the program, which does not have access to this information. As
such, it was not possible to obtain the frequency of people coming forward in response
to the reminder for this reason. It would be informative to know whether those coming
for the test in response to the reminder had experienced a recent bowel symptom and
whether this played a role in their motivation for making an appointment when previously
they had not attended. These individuals may be more likely to have bowel disease
and thereby benefit from the test. Future studies using questionnaires alongside these
reminders might be able to provide more information on this issue. Finally, as the
reminder used in this study consisted of multiple components, including a choice of
practitioner and options for the time and day of the appointment, it is not clear
how much each component contributed to uptake. It would be prudent to evaluate each
component independently and to investigate whether additional components, such as
general practice endorsement, having the enema administered at the hospital, and offering
a pre-booked appointment, may augment the observed effect [41].
Consistent with previous studies, the present trial highlights that interventions
which target nonresponders to take up screening after a missed appointment can be
effective and enhanced in a number of ways [41]. The study is the first to show that sending a reminder to former nonattenders is
also effective and that these individuals should receive such reminders alongside
former nonresponders. Additional reminders, possibly delivered at 24, 36, and even
48 months might improve uptake even further [27]
[28]. As an extension of the present trial, we are currently investigating whether there
is an added benefit to sending a second reminder 24 months after the initial invitation.
Conclusions
Use of a 12-month reminder in a single center of the English BSS Programme was effective
and improved uptake among former nonresponders and nonattenders. Inclusion of a theory-based
leaflet added significantly to this strategy, improving uptake even further. It is
important now to test this strategy across multiple centers and the wider population.
If consistent with the current study, implementing a reminder would increase population
coverage and concomitantly increase the number of CRCs prevented by the program.
