Does the Wording of Text Message Reminders Improve Uptake in Breast Screening?

Background

In the UK breast cancer is the most common female cancer affecting 1 in 8 women, accounting for 31% of all cancers and 15% of cancer deaths in women.1 The NHS breast screening programme (NHSBSP) offers 3-yearly screening to women aged 50-70 years. A pilot of extended screening ages of 47-73 is being trialed in some geographical areas.

In 2015, 75% of women in England attended breast screening. In the same year only 68.3% of women in London and 65.9% in West London respectively had an up-to-day mammogram in the previous 3 years. 2 This falls below both the national average and target uptake of 70%. The success of a screening programme not only relies on the sensitivity of the screening tool and availability of successful treatment but also on reaching the asymptomatic but at risk population.

Recent evidence in breast screening has shown that a simple text message reminder (TMR) is effective. 3,4. One randomized controlled trial showed an increase in attendance by as much as 12%.5 As a result TMRs are being used more frequently by NHS organisations, but with little consideration paid their word content.

Further research testing the word content of TMR to remind patients of routine outpatient hospital appointments has shown that the word choice matters. The most effective message reduced 'did not attend' rates by 3.4 percentage points compared to a control TMR at a London Hospital. 6

Objective

This study aims to test if the word content of TMRs affects the uptake of breast screening.

Trial design

The investigators propose to collaborate with West of London Breast Screening Service (WoLBSS) to test different TMRs through an RCT to increase screening uptake.

The trial will test the effect of differently worded TMRs on the uptake of breast screening in eligible women in West London. This three-armed RCT will test two intervention arms against the current control arm.

Current practice is for women to receive two TMRs at 7 and 4 days prior to their appointment. Therefore the current message will be maintained as the control message for both TMR 1 (7 days prior to appointment) and TMR 2 (4 days prior to appointment). The intervention arms will contain the intervention TMRs and will be sent out as both TMR1 and TMR2. During the trial period women due for screening will be randomized to receive one of the three trial arms (control, intervention arm 1 or intervention arm 2. The TMRs will be sent out 7 days and 4 days before the breast screening appointment during the trial period as they are currently sent out as part of routine practice. Women will receive the same message on the 7th day and the 4th day before their appointment. WoLBSS will inform the research team of a woman's attendance status at the end of the screening rounds.

Methods:

Study setting

This study will be run in West London in collaboration with WoLBSS, which currently provides the routine TMR service based on data from a previous trial.5 The research team will work closely with the WoLBSS team to run the RCT through the existing infrastructure used to send out current messages.

Eligibility Criteria

Inclusion Criteria

Women who are eligible and due for breast screening in West London.

Exclusion Criteria

Women who are not eligible for breast screening due to for example bilateral mastectomies.

Women who have chosen to not have screening and have informed the screening centre or their GP.

Withdrawal Criteria

Any woman who chooses to not take part in screening or to receive a TMR and has informed the screening centre or their GP.

Interventions

The RCT will have three arms.

1. control (current TMR)
2. Intervention 1 - Behavioural regulation TMR
3. Intervention 2 - Priority TMR

The choice of the content of the intervention text messages has been based on a survey of 1000 women run by the research team that evaluated how 15 behavioural constructs predicted the outcome measures of women's history of attendance and their intention to attend. Two strong constructs that predict both history of attendance and intention to attend were 'behavioural regulation' i.e. making a plan/note in a diary to attend breast screening and 'priority' i.e. how much a woman feels that screening is a priority.

Therefore the word content of two TMRs was constructed to take these into account.

Both intervention TMRs have been reviewed and approved by the WoLBSS routine care team as well as the clinical lead for screening and a Public Health Consultant in Cancer screening at NHS England.

Control - (current TMR)

'Don't forget your breast screening appointment at (time) (date) at (Site). Please call 02033136644 to rearrange or cancel'

Intervention arm 1 - Behavioural Regulation

'Don't forget to make a diary note of your screening mammogram appt at (time) (date) at (site). Call 02033136644 to rearrange.'

Intervention arm 2 - Priority

'Is your breast screening appointment on your priority list? Apt at (time), (date) at (site). Call 02033136644 to rearrange.'

Outcomes

The primary end point is the rate of uptake of screening at the end of the screening round. Trial arms will be powered and sample size collected to detect a 3% uptake in screening. The trial will close when the sample size is reached.

The secondary outcomes are uptake according to age, deprivation score and previous history of attendance.

Participant Timeline

Women will be routinely invited for their screening mammogram as per standard procedure approximately three weeks before their appointment. The research team will then randomise an anonymized list of women to receive one of the three TMRs. The TMR will then be sent at the same times as the current standard TMRs are sent out 7 and 4 days prior to the appointment.

Sample Size

The sample size was calculated to detect a 3% difference in uptake between groups with a 5% margin for type I and 20% margin for type II error. This expected improvement in uptake is deemed to be both realistic and operationally significant to the service. The uptake in West London in 2015 was 65.9%. The investigators therefore calculated a minimum sample size for each trial arm to be 3,832 with a total sample size of 11,496.

Recruitment

For each screening round WoLBSS will invite women for breast screening as per usual. It is not possible to identify at the time of randomization which women have provided WoLBSS with their phone number. For this reason all women due for screening will be randomised. However, only women who have given their mobile phone number will be able to receive a TMR.

Randomisation

Each screening round a list of women will be forward to iPlato Patient Care Messaging with their allocated appointment date, time and location. The participant list will then be randomised by iPlato into trial arms and allocated to receive one of the three TMRs using the Mersenne Twister algorithm.

Once randomised the list of women due for screening will be uploaded to the iPlato platform. The iPlato system will link each patients NHS number to WoLBSS record and retrieve a mobile phone number where one is available.

WoLBSS uses iPlato Patient Care Messaging currently to send out the existing text message reminder. iPlato is a communication platform developed specifically for health services and can send reminder messages on behalf of GPs and hospitals.

Data collection and management

Data will be collected by the WoLBSS IT system including the primary research outcome of attendance status. iPlato will collect data on accuracy of phone numbers and messages delivery status. Anonymised data will be retrieved and passed onto the research team. The data will be handled on NHS N3 computers with password protection, the research team's personal laptops with secure password protection and Imperial College London's secure virtual private network. Any transmission of data via email will be within secure NHS.net and NHS trust email accounts between WoLBSS and the research team and within the secure Imperial College email accounts and the virtual private network amongst the research team.

Data collected will include the age, year of birth, first 3-4 digits of the postcode, ethnicity, previous attendance history and availability of a mobile phone number.

It is possible that participants are subject to other trials however due to the nature of the trial it is not appropriate to contact patients to find out. It is not necessary in terms of impact on the results.

Consent

The premise of the study is whether a subtle change in a text message can increase the number of patients attending their breast screening appointment. A pre-warning that patients may receive TMR that may be modified would undermine the premise of the study and affect results. This is because of participation bias, which would indicate that subjects who know they are participating in a trial are more likely to try and do what they think the researcher wants them to do. Taking explicit consent would therefore also undermine the study design.

Only women who have given them mobile phone number to WoLBSS directly, will be able to receive a TMR prior to their appointment. Therefore they have in the past given explicit consent to receive a TMR by the service. No woman who has not previously given her mobile number to the service will receive a TMR.